Task¶

You are provided with registration data of 2298 mobile subscribers at base stations within one month.

The data format is as follows:

lac: Location Area Code (LAC). The Location Area Code, abbreviated as LAC is the unique number given to each location area within the network. The served area of a cellular radio access network is usually divided into location areas, consisting of one or several radio cells.
cid: Cell ID (CID). A GSM Cell ID (CID) is a generally unique number used to identify each base transceiver station (BTS) or sector of a BTS within a location area code (LAC) if not within a GSM network.
ts: UNIX-time. Means the number of seconds elapsed from 01.01.1970 to the event in question.
fulldate: Date
hash_id: Mobile user ID

The data is divided into two parts by the fulldate field:

from the 1st to the 14th of the month;
from the 16th to the 30th of the month.

Each part uses its own system for assigning identifiers to subscribers, which leads to the fact that the same subscriber in different parts of the sample has different identifiers: from 1st to 14th - one, from 16th to 30th - another.

Develop an algorithm that will establish a unique correspondence between identifiers from the first and second parts of the sample.

The expected result is a table of correspondences of the form id1 - id2 in *.csv format, where id1 is an identifier from the first part of the selection, id2 is an identifier from the second part of the selection.
To help you, a standard of 491 pairs of id1 and id2 is provided. You need to determine the remaining 1000+ matches.

Imports¶

In [1]:

# !pip install category_encoders eli5

In [2]:

import os
import sys
import warnings

import eli5
import numpy as np
import pandas as pd
import seaborn as sns
from sklearn.exceptions import ConvergenceWarning
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import precision_score, recall_score
from sklearn.model_selection import train_test_split
from tqdm.notebook import tqdm

In [3]:

from utils.memory_utils import reduce_mem_usage
from utils.time_utils import get_time_features, time_features
from utils.train_utils import fit_and_score, tf_idf_transform

In [4]:

sns.set(rc={"figure.dpi": 100, "savefig.dpi": 300})
sns.set_context("notebook")
sns.set_style("ticks")

%matplotlib inline
%config InlineBackend.figure_format = 'retina'

tqdm.pandas()

In [5]:

ConvergenceWarning("ignore")

warnings.filterwarnings("ignore")
if not sys.warnoptions:
    warnings.simplefilter("ignore")
    os.environ[
        "PYTHONWARNINGS"
    ] = "ignore::UserWarning,ignore::ConvergenceWarning,ignore::RuntimeWarning"

In [6]:

SEED = 42

Paths¶

Input¶

In [7]:

RELATIVE_PATH = "../data/"

In [8]:

DATA_PATH_PARQUET = os.path.join(RELATIVE_PATH, "01_data.parquet")
REFERENCE_PATH_PARQUET = os.path.join(RELATIVE_PATH, "02_etalon.parquet")

Output¶

In [9]:

FINAL_MAPPING_PATH_CSV = os.path.join(RELATIVE_PATH, "final_mapping.csv")

Loading Data¶

I've re-saved files in parquet format for faster loading (6x) and less memory consumption while storing (5x) (notebook 01_optimizing_storage.ipynb)

In [10]:

data = pd.read_parquet(DATA_PATH_PARQUET)
data.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 9153692 entries, 0 to 9153691
Data columns (total 5 columns):
 #   Column    Dtype  
---  ------    -----  
 0   lac       int64  
 1   cid       int64  
 2   ts        float64
 3   fulldate  object 
 4   hash_id   int64  
dtypes: float64(1), int64(3), object(1)
memory usage: 349.2+ MB

In [11]:

reference = pd.read_parquet(REFERENCE_PATH_PARQUET)
reference.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 491 entries, 0 to 490
Data columns (total 2 columns):
 #   Column  Non-Null Count  Dtype
---  ------  --------------  -----
 0   id1     491 non-null    int64
 1   id2     491 non-null    int64
dtypes: int64(2)
memory usage: 7.8 KB

Dataset Overview¶

In [12]:

data.sample(n=10, random_state=SEED)

Out[12]:

	lac	cid	ts	fulldate	hash_id
5020211	9732	198732291	1.538248e+09	2018-09-29	2193027
1751775	9731	198722053	1.536042e+09	2018-09-04	1167117
6373931	9934	197611523	1.537859e+09	2018-09-25	2129824
7344952	9745	10301	1.537101e+09	2018-09-16	2893856
5139826	9713	198838786	1.537440e+09	2018-09-20	2677662
8988177	7716	60325	1.537241e+09	2018-09-18	2862389
7552833	9716	198688516	1.537810e+09	2018-09-24	2521886
4635300	9937	249030668	1.537884e+09	2018-09-25	2571471
64115	9716	31422	1.536637e+09	2018-09-11	1213812
6502056	9785	198067460	1.537098e+09	2018-09-16	2841115

Dataset Close Look¶

In this section, let's look at what columns we have, what type they are, how many non-zero values, etc.

Reference¶

In [13]:

reference.plot()

Out[13]:

<AxesSubplot:>

In [14]:

sns.pairplot(reference)

Out[14]:

<seaborn.axisgrid.PairGrid at 0x1f32b9eefa0>

In [15]:

reference.hist(figsize=(15, 5), bins=100)

Out[15]:

array([[<AxesSubplot:title={'center':'id1'}>,
        <AxesSubplot:title={'center':'id2'}>]], dtype=object)

We will not be able to find any dependencies in reference's data

Data¶

In [16]:

data.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 9153692 entries, 0 to 9153691
Data columns (total 5 columns):
 #   Column    Dtype  
---  ------    -----  
 0   lac       int64  
 1   cid       int64  
 2   ts        float64
 3   fulldate  object 
 4   hash_id   int64  
dtypes: float64(1), int64(3), object(1)
memory usage: 349.2+ MB

Fields:

lac: Location Area Code (LAC). The Location Area Code, abbreviated as LAC is the unique number given to each location area within the network. The served area of a cellular radio access network is usually divided into location areas, consisting of one or several radio cells.
cid: Cell ID (CID). A GSM Cell ID (CID) is a generally unique number used to identify each base transceiver station (BTS) or sector of a BTS within a location area code (LAC) if not within a GSM network.
ts: UNIX-time. Means the number of seconds elapsed from 01.01.1970 to the event in question.
fulldate: Date
hash_id: Mobile user ID

`lac`¶

In [17]:

data["lac"].describe(
    percentiles=[0.01, 0.05, 0.25, 0.5, 0.75, 0.95, 0.99, 0.999]
).apply(lambda x: format(x, "f"))

Out[17]:

count    9153692.000000
mean        8950.529107
std         1377.873462
min         3302.000000
1%          5015.000000
5%          5061.000000
25%         7752.000000
50%         9716.000000
75%         9772.000000
95%         9935.000000
99%         9942.000000
99.9%       9976.000000
max        43000.000000
Name: lac, dtype: object

In [18]:

data["lac"].hist(bins=30)

Out[18]:

<AxesSubplot:>

In [19]:

data[data["lac"] <= 9976]["lac"].hist(bins=100)

Out[19]:

<AxesSubplot:>

There are only a few popular ranges of values for lac

In [20]:

data["lac"].value_counts()

Out[20]:

9716     951184
9734     567379
7716     537042
9916     534000
9934     490347
          ...  
23131         1
23120         1
9109          1
9560          1
9101          1
Name: lac, Length: 364, dtype: int64

We can see that we have 364 unique locations, although some of them are not very popular, while some of them are.

There are some outliers

In [21]:

data[data["lac"] > 9976]["lac"].value_counts()

Out[21]:

41800    31
43000    22
41801    19
23130     4
23111     3
23141     2
23120     1
23131     1
Name: lac, dtype: int64

`cid`¶

In [22]:

data["cid"].value_counts()

Out[22]:

3323         33830
33104        24805
3253         23217
33102        21505
14642        19027
             ...  
61225            1
36645            1
231806230        1
197768708        1
28596            1
Name: cid, Length: 75052, dtype: int64

We have a lot of unique values

In [23]:

data["cid"].hist()

Out[23]:

<AxesSubplot:>

In [24]:

data["cid"].describe(
    percentiles=[0.01, 0.05, 0.25, 0.36, 0.3694, 0.3695, 0.5, 0.75, 0.95, 0.99, 0.999]
).apply(lambda x: format(x, "f"))

Out[24]:

count       9153692.000000
mean      110670504.759660
std        98943536.660267
min               1.000000
1%              517.000000
5%             2317.000000
25%           32048.000000
36%           62477.000000
36.94%        65522.000000
36.95%       858479.000000
50%       129608451.000000
75%       198014732.000000
95%       232057857.000000
99%       253467908.000000
99.9%     254263301.000000
max       254855684.000000
Name: cid, dtype: object

Strange difference between 36.94 and 36.95 percentile. Let's find out more about cid

Fixing CID¶

Long cell ID vs. short cell ID
The formula for the long cell ID is as follows:
Long CID = 65536 * RNC + CID

RNC: the Radio Network Controller
CID: a short cell ID (an integer in the range of 0 to 65535)

If you have the Long CID, you can get RNC and CID in the following way:
RNC = Long CID / 65536 (integer division)
CID = Long CID mod 65536 (modulo operation)

Example for long cell ID 66808694:
RNC = 66808694 / 65536 = 1019
CID = 66808694 mod 65536 = 27510

Aha. We have Long and Short CID mixed. Let's make all Long CID Short.

In [25]:

data["cid"] = data["cid"].progress_apply(
    lambda cid: cid % 65536 if cid > 65535 else cid
)

  0%|          | 0/9153692 [00:00<?, ?it/s]

In [26]:

data["cid"].describe().apply(lambda x: format(x, "f"))

Out[26]:

count    9153692.000000
mean       28729.860331
std        19928.071245
min            1.000000
25%        10273.000000
50%        28965.000000
75%        45571.000000
max        65534.000000
Name: cid, dtype: object

In [27]:

data["cid"].value_counts()

Out[27]:

3323     33830
25861    26781
33104    24805
30466    23887
3253     23217
         ...  
42709        1
61044        1
23468        1
15264        1
27058        1
Name: cid, Length: 38926, dtype: int64

Now it looks better

In [28]:

data["cid"].hist()

Out[28]:

<AxesSubplot:>

`lac` and `cid` interaction¶

In [29]:

lac_cid_group = (
    pd.DataFrame(data.groupby(["lac", "cid"])["cid"].count())
    .rename(columns={"cid": "count"})
    .reset_index()
)

In [30]:

with pd.option_context("display.max_rows", None, "display.max_columns", None):
    display(lac_cid_group.sort_values(by=["cid", "lac"]).iloc[:20])

	lac	cid	count
3	5000	1	27
10986	5029	1	16
13378	5037	1	2
15882	5047	1	33
46517	9000	1	71
46681	9001	1	1
46872	9002	1	1
53435	9054	1	25
54197	9071	1	35
57540	9704	1	17
59837	9715	1	22
63461	9733	1	6
67446	9746	1	6
68400	9748	1	8
68937	9751	1	112
73291	9773	1	7
75695	9790	1	280
76526	9795	1	28
78712	9915	1	30
81516	9950	1	6

The same cid can belong to different lac, which is strange. But it can be fine, if for every lac there is its own cid

In [31]:

lac_cid_group.shape

Out[31]:

(82191, 3)

There are 82191 unique combinations of lac and cid

Let's create a feature, where lac and cid will be together

In [32]:

data["lac_cid"] = data["lac"].astype("str") + "_" + data["cid"].astype("str")

`ts`¶

ts column will be transformed into datetime format in the next section

`fulldate`¶

fulldate column will be transformed into datetime format in the next section

`hash_id`¶

In [33]:

unique_hash_ids = data["hash_id"].unique()

In [34]:

data["hash_id"].value_counts()

Out[34]:

1088711    34138
2187102    30128
2674156    18115
2102524    13183
1819111    12862
           ...  
1968066        4
1255062        4
1754876        3
1534909        2
1570683        1
Name: hash_id, Length: 4542, dtype: int64

We know that there are reference out of $491$ pairs, which makes $982$ IDs that have direct matching

We also know that there are $1000+$ more pairs.

Maybe we have IDs that don't have a match.

Data transformation¶

Saving a bit of memory¶

As soon as our dataset is not optimally stored, we can optimize memory consumption.

In [35]:

fulldate = data["fulldate"]
data = reduce_mem_usage(data.drop("fulldate", axis=1))
data["fulldate"] = fulldate

Memory usage of dataframe is 349.19 MB

  0%|          | 0/5 [00:00<?, ?it/s]

Memory usage after optimization is: 177.24 MB
Decreased by 49.2%

`ts` and `fulldate`¶

Transforming `ts` and `fulldate` into `datetime` format¶

We know that ts column is the number of seconds elapsed from 01.01.1970 to the event.
Let's encode it to datetime format.

In [36]:

data.insert(3, "datetime_ts", pd.to_datetime(data["ts"], unit="s"))
data.drop("ts", axis=1, inplace=True, errors="ignore")
data["fulldate"] = pd.to_datetime(data["fulldate"])

In [37]:

data.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 9153692 entries, 0 to 9153691
Data columns (total 6 columns):
 #   Column       Dtype         
---  ------       -----         
 0   lac          int32         
 1   cid          int32         
 2   datetime_ts  datetime64[ns]
 3   hash_id      int32         
 4   lac_cid      category      
 5   fulldate     datetime64[ns]
dtypes: category(1), datetime64[ns](2), int32(3)
memory usage: 282.0 MB

Date and Time features extraction¶

In [38]:

datetime_features, datetime_feature_names = get_time_features(
    data["fulldate"], data["datetime_ts"]
)

In [39]:

datetime_features.head()

Out[39]:

	hour	morning	weekday	weekend
0	7	1	4	0
1	7	1	5	0
2	9	1	2	0
3	7	1	7	1
4	9	1	3	0

In [40]:

data = pd.concat([data, datetime_features], axis=1)

Checking data integrity¶

In [41]:

data[(data["fulldate"].dt.day == 15) | (data["fulldate"].dt.day == 31)]

Out[41]:

	lac	cid	datetime_ts	hash_id	lac_cid	fulldate	hour	morning	day	evening	night	weekday	weekend

As promised, the data was split into two parts by day:

from the $1^{st}$ to the $14^{th}$ of the month
from the $16^{th}$ to the $30^{th}$ of the month

So there are no $15$ and $31$ days in the sample. It's fine.

In [42]:

sorted(data["fulldate"].dt.date.unique())

Out[42]:

[datetime.date(2018, 9, 1),
 datetime.date(2018, 9, 2),
 datetime.date(2018, 9, 3),
 datetime.date(2018, 9, 4),
 datetime.date(2018, 9, 5),
 datetime.date(2018, 9, 6),
 datetime.date(2018, 9, 7),
 datetime.date(2018, 9, 8),
 datetime.date(2018, 9, 9),
 datetime.date(2018, 9, 10),
 datetime.date(2018, 9, 11),
 datetime.date(2018, 9, 12),
 datetime.date(2018, 9, 13),
 datetime.date(2018, 9, 14),
 datetime.date(2018, 9, 16),
 datetime.date(2018, 9, 17),
 datetime.date(2018, 9, 18),
 datetime.date(2018, 9, 19),
 datetime.date(2018, 9, 20),
 datetime.date(2018, 9, 21),
 datetime.date(2018, 9, 22),
 datetime.date(2018, 9, 23),
 datetime.date(2018, 9, 24),
 datetime.date(2018, 9, 25),
 datetime.date(2018, 9, 26),
 datetime.date(2018, 9, 27),
 datetime.date(2018, 9, 28),
 datetime.date(2018, 9, 29),
 datetime.date(2018, 9, 30)]

We have one full month - $\text{September 2018}$, except $15$ and $31$ days, as stated before.

There are $\text{61 297}$ records for which date from fulldate column doesn't match the date, which was obtained from ts column.
Not fine!

In [43]:

data[data["fulldate"].dt.date != data["datetime_ts"].dt.date]

Out[43]:

	lac	cid	datetime_ts	hash_id	lac_cid	fulldate	hour	morning	day	evening	night	weekday	weekend
0	7755	35950	2018-09-08 07:10:56	1361396	7755_35950	2018-09-13	7	1	0	0	0	4	0
1	7755	35950	2018-09-08 07:10:56	1361396	7755_35950	2018-09-14	7	1	0	0	0	5	0
2	7752	19203	2018-09-03 09:31:44	1361396	7752_19203	2018-09-04	9	1	0	0	0	2	0
3	7755	35950	2018-09-08 07:10:56	1361396	7755_35950	2018-09-09	7	1	0	0	0	7	1
4	7752	19203	2018-09-03 09:31:44	1361396	7752_19203	2018-09-05	9	1	0	0	0	3	0
...	...	...	...	...	...	...	...	...	...	...	...	...	...
9153681	7746	6551	2018-09-24 16:14:56	2241318	7746_6551	2018-09-26	16	0	1	0	0	3	0
9153684	7716	11870	2018-09-13 18:22:56	2241318	7716_11870	2018-09-16	18	0	1	0	0	7	1
9153685	7716	11870	2018-09-13 18:22:56	2241318	7716_11870	2018-09-19	18	0	1	0	0	3	0
9153686	7746	6551	2018-09-24 16:14:56	2241318	7746_6551	2018-09-28	16	0	1	0	0	5	0
9153691	7716	11870	2018-09-13 18:22:56	2241318	7716_11870	2018-09-17	18	0	1	0	0	1	0

61297 rows × 13 columns

And there are some records for which splitting in two groups could be ambiguous.
For example, fulldate is 2018-09-19 and datetime_ts is 2018-09-13. Using one column data should be in one partition and using other column data should be in another partition.

User journeys creation from `lac_cid` feature¶

I've found out that we can create a list of lac_cid values for every user in sorted order (by datetime_ts column).
It will show how every specific user was registered on different base stations along the way of time.

For example, let's take one hash_id

In [44]:

unique_hash_ids[0]

Out[44]:

And look at the path for user $1361396$

In [45]:

" ".join(
    data[data["hash_id"] == 1361396].sort_values(by="datetime_ts")["lac_cid"].values
)

Out[45]:

'5007_7088 7752_19207 7752_19203 7752_19203 7752_19203 7752_19207 7752_19207 7752_19203 7755_23103 7755_23103 7755_23103 7755_35950 7755_35950 7755_35950 7755_35950 7755_35950 7755_35950 7755_35950'

Let's do this for every user in our dataset.
We can save the whole path for every user, but probably we will be good with only part of it.

The length of the path could be some arbitrary number like 100 or 200, or it could be some statistic, like mean/median/specific percentile of the length distribution across users.

For now, let's take median value for the dataset.

In [46]:

user_journeys = {}

In [47]:

n_journeys = int(data.groupby("hash_id")["lac_cid"].count().median())
n_journeys

Out[47]:

In [48]:

for user in tqdm(unique_hash_ids, desc="Creating user journeys"):
    user_journeys[user] = " ".join(
        data[data["hash_id"] == user]
        .sort_values(by="datetime_ts")["lac_cid"]
        .values[:n_journeys]
    )

Creating user journeys:   0%|          | 0/4542 [00:00<?, ?it/s]

Of course, we'll create a column based on this data.

In [49]:

data["user_journey"] = data["hash_id"].map(user_journeys)

Datetime based features¶

Currently, these features are not used in the solution.

In [50]:

data.sample(10, random_state=SEED)

Out[50]:

	lac	cid	datetime_ts	hash_id	lac_cid	fulldate	hour	morning	day	evening	night	weekday	weekend	user_journey
5020211	9732	27139	2018-09-29 19:07:44	2193027	9732_27139	2018-09-29	19	0	0	1	0	6	1	9732_27139 9732_27168 9732_27139 9732_27168 97...
1751775	9731	16901	2018-09-04 06:19:44	1167117	9731_16901	2018-09-04	6	0	0	0	1	2	0	9091_52008 9060_16643 9091_52008 5091_64312 90...
6373931	9934	20483	2018-09-25 07:00:16	2129824	9934_20483	2018-09-25	7	1	0	0	0	2	0	7716_38835 9716_31069 7716_48445 9716_61958 77...
7344952	9745	10301	2018-09-16 12:33:04	2893856	9745_10301	2018-09-16	12	0	1	0	0	7	1	7716_3323 7716_9553 9716_34606 7716_3323 7716_...
5139826	9713	2562	2018-09-20 10:35:44	2677662	9713_2562	2018-09-20	10	1	0	0	0	4	0	9716_30230 9716_17210 9716_3145 7716_3155 9716...
8988177	7716	60325	2018-09-18 03:22:40	2862389	7716_60325	2018-09-18	3	0	0	0	1	2	0	7716_60325 7716_9553 7716_60325 7716_9553 7716...
7552833	9716	48900	2018-09-24 17:27:28	2521886	9716_48900	2018-09-24	17	0	1	0	0	1	0	7730_460 7716_2147 9716_30220 9716_36733 7730_...
4635300	9937	59404	2018-09-25 13:54:08	2571471	9937_59404	2018-09-25	13	0	1	0	0	2	0	9758_23298 9734_14852 9758_23298 9734_14849 97...
64115	9716	31422	2018-09-11 03:33:20	1213812	9716_31422	2018-09-11	3	0	0	0	1	2	0	7716_60475 9716_31422 7716_60475 9716_31422 77...
6502056	9785	17668	2018-09-16 11:46:08	2841115	9785_17668	2018-09-16	11	1	0	0	0	7	1	9716_9218 9716_9218 7716_3579 7716_3579 9716_3...

In [51]:

hour_tf = time_features(data, column="hour", prefix="reg_hour", add_total_count=True)
hour_tf

Out[51]:

hour	reg_hour_0	reg_hour_1	reg_hour_2	reg_hour_3	reg_hour_4	reg_hour_5	reg_hour_6	reg_hour_7	reg_hour_8	reg_hour_9	...	reg_hour_15	reg_hour_16	reg_hour_17	reg_hour_18	reg_hour_19	reg_hour_20	reg_hour_21	reg_hour_22	reg_hour_23	reg_hour_total
hash_id
1000773	0.000000	0.000000	0.017241	0.025862	0.008621	0.017241	0.034483	0.077586	0.051724	0.077586	...	0.051724	0.068966	0.077586	0.060345	0.068966	0.094828	0.008621	0.000000	0.000000	116.0
1000808	0.041284	0.041284	0.041284	0.041284	0.050459	0.055046	0.050459	0.013761	0.018349	0.032110	...	0.055046	0.059633	0.045872	0.055046	0.059633	0.059633	0.059633	0.055046	0.050459	218.0
1000868	0.028777	0.028777	0.032374	0.035971	0.035971	0.043165	0.046763	0.039568	0.039568	0.046763	...	0.043165	0.050360	0.046763	0.050360	0.046763	0.046763	0.043165	0.039568	0.050360	278.0
1001103	0.027027	0.000000	0.000000	0.000000	0.081081	0.081081	0.081081	0.081081	0.081081	0.081081	...	0.054054	0.054054	0.000000	0.000000	0.000000	0.000000	0.000000	0.054054	0.000000	37.0
1001168	0.043103	0.025862	0.017241	0.051724	0.051724	0.068966	0.025862	0.034483	0.017241	0.025862	...	0.043103	0.060345	0.051724	0.043103	0.025862	0.034483	0.051724	0.051724	0.060345	116.0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
2998617	0.000000	0.008475	0.008475	0.033898	0.042373	0.033898	0.050847	0.076271	0.076271	0.059322	...	0.008475	0.025424	0.042373	0.076271	0.076271	0.110169	0.050847	0.000000	0.000000	118.0
2999034	0.005405	0.005405	0.016216	0.054054	0.059459	0.054054	0.059459	0.070270	0.064865	0.059459	...	0.043243	0.059459	0.064865	0.048649	0.021622	0.010811	0.005405	0.005405	0.005405	185.0
2999248	0.008163	0.016327	0.008163	0.036735	0.053061	0.057143	0.053061	0.057143	0.057143	0.053061	...	0.053061	0.057143	0.053061	0.048980	0.044898	0.016327	0.028571	0.028571	0.008163	245.0
2999753	0.017045	0.011364	0.005682	0.017045	0.045455	0.051136	0.068182	0.056818	0.028409	0.034091	...	0.062500	0.068182	0.045455	0.062500	0.039773	0.062500	0.034091	0.028409	0.022727	176.0
2999841	0.000000	0.000000	0.000000	0.000000	0.021739	0.108696	0.043478	0.021739	0.043478	0.086957	...	0.086957	0.065217	0.021739	0.152174	0.108696	0.000000	0.000000	0.000000	0.000000	46.0

4542 rows × 25 columns

In [52]:

weekday_tf = time_features(
    data, column="weekday", prefix="reg_weekday", add_total_count=True
)
hour_tf

Out[52]:

hour	reg_hour_0	reg_hour_1	reg_hour_2	reg_hour_3	reg_hour_4	reg_hour_5	reg_hour_6	reg_hour_7	reg_hour_8	reg_hour_9	...	reg_hour_15	reg_hour_16	reg_hour_17	reg_hour_18	reg_hour_19	reg_hour_20	reg_hour_21	reg_hour_22	reg_hour_23	reg_hour_total
hash_id
1000773	0.000000	0.000000	0.017241	0.025862	0.008621	0.017241	0.034483	0.077586	0.051724	0.077586	...	0.051724	0.068966	0.077586	0.060345	0.068966	0.094828	0.008621	0.000000	0.000000	116.0
1000808	0.041284	0.041284	0.041284	0.041284	0.050459	0.055046	0.050459	0.013761	0.018349	0.032110	...	0.055046	0.059633	0.045872	0.055046	0.059633	0.059633	0.059633	0.055046	0.050459	218.0
1000868	0.028777	0.028777	0.032374	0.035971	0.035971	0.043165	0.046763	0.039568	0.039568	0.046763	...	0.043165	0.050360	0.046763	0.050360	0.046763	0.046763	0.043165	0.039568	0.050360	278.0
1001103	0.027027	0.000000	0.000000	0.000000	0.081081	0.081081	0.081081	0.081081	0.081081	0.081081	...	0.054054	0.054054	0.000000	0.000000	0.000000	0.000000	0.000000	0.054054	0.000000	37.0
1001168	0.043103	0.025862	0.017241	0.051724	0.051724	0.068966	0.025862	0.034483	0.017241	0.025862	...	0.043103	0.060345	0.051724	0.043103	0.025862	0.034483	0.051724	0.051724	0.060345	116.0
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
2998617	0.000000	0.008475	0.008475	0.033898	0.042373	0.033898	0.050847	0.076271	0.076271	0.059322	...	0.008475	0.025424	0.042373	0.076271	0.076271	0.110169	0.050847	0.000000	0.000000	118.0
2999034	0.005405	0.005405	0.016216	0.054054	0.059459	0.054054	0.059459	0.070270	0.064865	0.059459	...	0.043243	0.059459	0.064865	0.048649	0.021622	0.010811	0.005405	0.005405	0.005405	185.0
2999248	0.008163	0.016327	0.008163	0.036735	0.053061	0.057143	0.053061	0.057143	0.057143	0.053061	...	0.053061	0.057143	0.053061	0.048980	0.044898	0.016327	0.028571	0.028571	0.008163	245.0
2999753	0.017045	0.011364	0.005682	0.017045	0.045455	0.051136	0.068182	0.056818	0.028409	0.034091	...	0.062500	0.068182	0.045455	0.062500	0.039773	0.062500	0.034091	0.028409	0.022727	176.0
2999841	0.000000	0.000000	0.000000	0.000000	0.021739	0.108696	0.043478	0.021739	0.043478	0.086957	...	0.086957	0.065217	0.021739	0.152174	0.108696	0.000000	0.000000	0.000000	0.000000	46.0

4542 rows × 25 columns

In [53]:

data.groupby("hash_id")["lac"].nunique()

Out[53]:

hash_id
1000773    44
1000808    26
1000868    51
1001103    24
1001168    31
           ..
2998617    27
2999034    27
2999248    69
2999753    50
2999841    12
Name: lac, Length: 4542, dtype: int64

In [54]:

data.groupby("hash_id")["cid"].nunique()

Out[54]:

hash_id
1000773    381
1000808    170
1000868    499
1001103    145
1001168    130
          ... 
2998617    288
2999034    111
2999248    638
2999753    394
2999841     30
Name: cid, Length: 4542, dtype: int64

In [55]:

# data = data.join(tf, on="hash_id")

In [56]:

# data = data.join(lac_counts, on="hash_id")

In [57]:

# data = data.join(cid_counts, on="hash_id")

Splitting data¶

Splitting¶

Let's prepare hash_id's that will indicate that the row is in train dataset.

In [58]:

train_ind = reference.values.flatten()

assert len(train_ind) == len(np.unique(train_ind))
print(reference.shape)

(491, 2)

Training dataset¶

Slicing data to get training data

In [59]:

train_data = data.loc[data["hash_id"].isin(train_ind)]
train_data.shape

Out[59]:

(2019095, 14)

In [60]:

train_data_left = train_data.loc[
    train_data["hash_id"].isin(reference["id1"])
].reset_index(drop=True)
train_data_left.shape

Out[60]:

(983406, 14)

In [61]:

train_data_right = train_data.loc[
    train_data["hash_id"].isin(reference["id2"])
].reset_index(drop=True)
train_data_right.shape

Out[61]:

(1035689, 14)

Both train_data_left and train_data_right can serve as a training labelled dataset and as a test labelled dataset.
It can be reasonable to take right dataset, because it contains more data.

New ID for training dataset¶

Let's fix the problem, where the user has a different ID in different samples for training dataset

In [62]:

reference["new_id"] = list(range(reference.shape[0]))
reference.shape

Out[62]:

(491, 3)

In [63]:

reference.head()

Out[63]:

	id1	id2	new_id
0	1361396	2695335	0
1	1795864	2458905	1
2	1543059	2730453	2
3	1028066	2539971	3
4	1533076	2712514	4

Now we map the ID's to new ID, which will be the same across left and right training datasets for easier evaluation.

In [102]:

left_map = dict(zip(reference["id1"], reference["new_id"]))
right_map = dict(zip(reference["id2"], reference["new_id"]))

In [65]:

train_data_left["hash_id"] = train_data_left["hash_id"].map(left_map)
train_data_left.head()

Out[65]:

	lac	cid	datetime_ts	lac_cid	fulldate	hour	morning	weekday	weekend	user_journey
0	7755	35950	2018-09-08 07:10:56	7755_35950	2018-09-13	7	1	4	0	5007_7088 7752_19207 7752_19203 7752_19203 775...
1	7755	35950	2018-09-08 07:10:56	7755_35950	2018-09-14	7	1	5	0	5007_7088 7752_19207 7752_19203 7752_19203 775...
2	7752	19203	2018-09-03 09:31:44	7752_19203	2018-09-04	9	1	2	0	5007_7088 7752_19207 7752_19203 7752_19203 775...
3	7755	35950	2018-09-08 07:10:56	7755_35950	2018-09-09	7	1	7	1	5007_7088 7752_19207 7752_19203 7752_19203 775...
4	7752	19203	2018-09-03 09:31:44	7752_19203	2018-09-05	9	1	3	0	5007_7088 7752_19207 7752_19203 7752_19203 775...

In [66]:

train_data_right["hash_id"] = train_data_right["hash_id"].map(right_map)
train_data_right.head()

Out[66]:

	lac	cid	datetime_ts	lac_cid	fulldate	hour	morning	day	evening	weekday	user_journey
0	7755	23103	2018-09-25 20:05:20	7755_23103	2018-09-25	20	0	0	1	2	7755_23103 7755_23103 7755_23103 7755_23103 77...
1	7752	19203	2018-09-27 08:32:00	7752_19203	2018-09-27	8	1	0	0	4	7755_23103 7755_23103 7755_23103 7755_23103 77...
2	7755	23103	2018-09-22 15:51:28	7755_23103	2018-09-24	15	0	1	0	1	7755_23103 7755_23103 7755_23103 7755_23103 77...
3	7755	23103	2018-09-18 20:39:28	7755_23103	2018-09-18	20	0	0	1	2	7755_23103 7755_23103 7755_23103 7755_23103 77...
4	7745	1383	2018-09-25 08:08:32	7745_1383	2018-09-25	8	1	0	0	2	7755_23103 7755_23103 7755_23103 7755_23103 77...

In [67]:

assert train_data_left["hash_id"].nunique() == train_data_right["hash_id"].nunique()

We still have the same number of classes in left and train datasets

Testing dataset¶

Slicing to get testing data

In [68]:

test_data = data.loc[~data["hash_id"].isin(train_ind)]
test_data.shape

Out[68]:

(7134597, 14)

In [69]:

test_data_left = test_data.loc[
    test_data["fulldate"].dt.day.isin(list(range(1, 15)))
].reset_index(drop=True)
test_data_left.shape

Out[69]:

(3393052, 14)

In [70]:

test_data_right = test_data.loc[
    test_data["fulldate"].dt.day.isin(list(range(16, 31)))
].reset_index(drop=True)
test_data_right.shape

Out[70]:

(3741545, 14)

In [71]:

test_data_left["hash_id"].nunique(), test_data_right["hash_id"].nunique()

Out[71]:

(1757, 1803)

This time we have more unique classes in the right dataset

Modelling¶

Preparing data¶

We'll be dropping duplicates as long as we need only user journey which is unique for every user ID.

We'll be using right dataset for training as long as it has more data.

In [72]:

X_train = train_data_right.drop_duplicates(["hash_id", "user_journey"])
X_test = train_data_left.drop_duplicates(["hash_id", "user_journey"])

y_train = X_train["hash_id"].values
y_test = X_test["hash_id"].values
train_journeys = X_train["user_journey"].values
test_journeys = X_test["user_journey"].values

TF-IDF for user journeys¶

These are untuned parameters for TF-IDF with which I've started

In [73]:

vectorizer_params = {
    "ngram_range": (1, 3),
    "max_features": 50_000,
    "tokenizer": lambda s: s.split(),
}

In [74]:

vectorizer = TfidfVectorizer(**vectorizer_params)
X_train_tfidf = vectorizer.fit_transform(train_journeys)

In [75]:

X_test_tfidf = vectorizer.transform(test_journeys)

Logistic Regression¶

In [76]:

clf = LogisticRegression(
    C=1,
    random_state=SEED,
    n_jobs=-1,
    verbose=2,
    multi_class="multinomial",  # to solve the task like Softmax Classifier
    solver="sag",  # for faster training
)

In [77]:

clf.fit(X_train_tfidf, y_train)

[Parallel(n_jobs=-1)]: Using backend ThreadingBackend with 16 concurrent workers.

convergence after 13 epochs took 7 seconds

[Parallel(n_jobs=-1)]: Done   1 out of   1 | elapsed:    7.3s finished

Out[77]:

LogisticRegression(C=1, multi_class='multinomial', n_jobs=-1, random_state=42,
                   solver='sag', verbose=2)

In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.

In [78]:

y_pred = clf.predict(X_test_tfidf)

In [79]:

precision_score(y_test, y_pred, average="macro")

Out[79]:

0.8339103869653768

In [80]:

recall_score(y_test, y_pred, average="macro")

Out[80]:

0.879837067209776

Pretty decent performance, huh?

I've tried another classifiers like KNN, Extreme Random Trees, Random Forest, but their performance is nowhere near as good as LogReg's performance.

Results analysis¶

In [81]:

eli5.show_weights(estimator=clf, feature_names=vectorizer.get_feature_names(), top=10)

Out[81]:

y=0 top features

y=1 top features

y=2 top features

y=3 top features

y=4 top features

y=5 top features

y=6 top features

y=7 top features

y=8 top features

y=9 top features

y=10 top features

y=11 top features

y=12 top features

y=13 top features

y=14 top features

y=15 top features

y=16 top features

y=17 top features

y=18 top features

y=19 top features

y=20 top features

y=21 top features

y=22 top features

y=23 top features

y=24 top features

y=25 top features

y=26 top features

y=27 top features

y=28 top features

y=29 top features

y=30 top features

y=31 top features

y=32 top features

y=33 top features

y=34 top features

y=35 top features

y=36 top features

y=37 top features

y=38 top features

y=39 top features

y=40 top features

y=41 top features

y=42 top features

y=43 top features

y=44 top features

y=45 top features

y=46 top features

y=47 top features

y=48 top features

y=49 top features

y=50 top features

y=51 top features

y=52 top features

y=53 top features

y=54 top features

y=55 top features

y=56 top features

y=57 top features

y=58 top features

y=59 top features

y=60 top features

y=61 top features

y=62 top features

y=63 top features

y=64 top features

y=65 top features

y=66 top features

y=67 top features

y=68 top features

y=69 top features

y=70 top features

y=71 top features

y=72 top features

y=73 top features

y=74 top features

y=75 top features

y=76 top features

y=77 top features

y=78 top features

y=79 top features

y=80 top features

y=81 top features

y=82 top features

y=83 top features

y=84 top features

y=85 top features

y=86 top features

y=87 top features

y=88 top features

y=89 top features

y=90 top features

y=91 top features

y=92 top features

y=93 top features

y=94 top features

y=95 top features

y=96 top features

y=97 top features

y=98 top features

y=99 top features

y=100 top features

y=101 top features

y=102 top features

y=103 top features

y=104 top features

y=105 top features

y=106 top features

y=107 top features

y=108 top features

y=109 top features

y=110 top features

y=111 top features

y=112 top features

y=113 top features

y=114 top features

y=115 top features

y=116 top features

y=117 top features

y=118 top features

y=119 top features

y=120 top features

y=121 top features

y=122 top features

y=123 top features

y=124 top features

y=125 top features

y=126 top features

y=127 top features

y=128 top features

y=129 top features

y=130 top features

y=131 top features

y=132 top features

y=133 top features

y=134 top features

y=135 top features

y=136 top features

y=137 top features

y=138 top features

y=139 top features

y=140 top features

y=141 top features

y=142 top features

y=143 top features

y=144 top features

y=145 top features

y=146 top features

y=147 top features

y=148 top features

y=149 top features

y=150 top features

y=151 top features

y=152 top features

y=153 top features

y=154 top features

y=155 top features

y=156 top features

y=157 top features

y=158 top features

y=159 top features

y=160 top features

y=161 top features

y=162 top features

y=163 top features

y=164 top features

y=165 top features

y=166 top features

y=167 top features

y=168 top features

y=169 top features

y=170 top features

y=171 top features

y=172 top features

y=173 top features

y=174 top features

y=175 top features

y=176 top features

y=177 top features

y=178 top features

y=179 top features

y=180 top features

y=181 top features

y=182 top features

y=183 top features

y=184 top features

y=185 top features

y=186 top features

y=187 top features

y=188 top features

y=189 top features

y=190 top features

y=191 top features

y=192 top features

y=193 top features

y=194 top features

y=195 top features

y=196 top features

y=197 top features

y=198 top features

y=199 top features

y=200 top features

y=201 top features

y=202 top features

y=203 top features

y=204 top features

y=205 top features

y=206 top features

y=207 top features

y=208 top features

y=209 top features

y=210 top features

y=211 top features

y=212 top features

y=213 top features

y=214 top features

y=215 top features

y=216 top features

y=217 top features

y=218 top features

y=219 top features

y=220 top features

y=221 top features

y=222 top features

y=223 top features

y=224 top features

y=225 top features

y=226 top features

y=227 top features

y=228 top features

y=229 top features

y=230 top features

y=231 top features

y=232 top features

y=233 top features

y=234 top features

y=235 top features

y=236 top features

y=237 top features

y=238 top features

y=239 top features

y=240 top features

y=241 top features

y=242 top features

y=243 top features

y=244 top features

y=245 top features

y=246 top features

y=247 top features

y=248 top features

y=249 top features

y=250 top features

y=251 top features

y=252 top features

y=253 top features

y=254 top features

y=255 top features

y=256 top features

y=257 top features

y=258 top features

y=259 top features

y=260 top features

y=261 top features

y=262 top features

y=263 top features

y=264 top features

y=265 top features

y=266 top features

y=267 top features

y=268 top features

y=269 top features

y=270 top features

y=271 top features

y=272 top features

y=273 top features

y=274 top features

y=275 top features

y=276 top features

y=277 top features

y=278 top features

y=279 top features

y=280 top features

y=281 top features

y=282 top features

y=283 top features

y=284 top features

y=285 top features

y=286 top features

y=287 top features

y=288 top features

y=289 top features

y=290 top features

y=291 top features

y=292 top features

y=293 top features

y=294 top features

y=295 top features

y=296 top features

y=297 top features

y=298 top features

y=299 top features

y=300 top features

y=301 top features

y=302 top features

y=303 top features

y=304 top features

y=305 top features

y=306 top features

y=307 top features

y=308 top features

y=309 top features

y=310 top features

y=311 top features

y=312 top features

y=313 top features

y=314 top features

y=315 top features

y=316 top features

y=317 top features

y=318 top features

y=319 top features

y=320 top features

y=321 top features

y=322 top features

y=323 top features

y=324 top features

y=325 top features

y=326 top features

y=327 top features

y=328 top features

y=329 top features

y=330 top features

y=331 top features

y=332 top features

y=333 top features

y=334 top features

y=335 top features

y=336 top features

y=337 top features

y=338 top features

y=339 top features

y=340 top features

y=341 top features

y=342 top features

y=343 top features

y=344 top features

y=345 top features

y=346 top features

y=347 top features

y=348 top features

y=349 top features

y=350 top features

y=351 top features

y=352 top features

y=353 top features

y=354 top features

y=355 top features

y=356 top features

y=357 top features

y=358 top features

y=359 top features

y=360 top features

y=361 top features

y=362 top features

y=363 top features

y=364 top features

y=365 top features

y=366 top features

y=367 top features

y=368 top features

y=369 top features

y=370 top features

y=371 top features

y=372 top features

y=373 top features

y=374 top features

y=375 top features

y=376 top features

y=377 top features

y=378 top features

y=379 top features

y=380 top features

y=381 top features

y=382 top features

y=383 top features

y=384 top features

y=385 top features

y=386 top features

y=387 top features

y=388 top features

y=389 top features

y=390 top features

y=391 top features

y=392 top features

y=393 top features

y=394 top features

y=395 top features

y=396 top features

y=397 top features

y=398 top features

y=399 top features

y=400 top features

y=401 top features

y=402 top features

y=403 top features

y=404 top features

y=405 top features

y=406 top features

y=407 top features

y=408 top features

y=409 top features

y=410 top features

y=411 top features

y=412 top features

y=413 top features

y=414 top features

y=415 top features

y=416 top features

y=417 top features

y=418 top features

y=419 top features

y=420 top features

y=421 top features

y=422 top features

y=423 top features

y=424 top features

y=425 top features

y=426 top features

y=427 top features

y=428 top features

y=429 top features

y=430 top features

y=431 top features

y=432 top features

y=433 top features

y=434 top features

y=435 top features

y=436 top features

y=437 top features

y=438 top features

y=439 top features

y=440 top features

y=441 top features

y=442 top features

y=443 top features

y=444 top features

y=445 top features

y=446 top features

y=447 top features

y=448 top features

y=449 top features

y=450 top features

y=451 top features

y=452 top features

y=453 top features

y=454 top features

y=455 top features

y=456 top features

y=457 top features

y=458 top features

y=459 top features

y=460 top features

y=461 top features

y=462 top features

y=463 top features

y=464 top features

y=465 top features

y=466 top features

y=467 top features

y=468 top features

y=469 top features

y=470 top features

y=471 top features

y=472 top features

y=473 top features

y=474 top features

y=475 top features

y=476 top features

y=477 top features

y=478 top features

y=479 top features

y=480 top features

y=481 top features

y=482 top features

y=483 top features

y=484 top features

y=485 top features

y=486 top features

y=487 top features

y=488 top features

y=489 top features

y=490 top features

Weight^?	Feature
+0.571	7752_19203
+0.510	7755_23103
+0.378	7755_23103 7755_23103
+0.362	7752_19203 7752_19203
+0.242	7755_23103 7752_19203
+0.242	7755_23103 7755_23103 7752_19203
+0.080	7745_1383
+0.057	7745_1383 9752_19207
+0.041	9752_19207
… 1 more positive …
… 49990 more negative …
-0.009	9752_14633

Weight^?	Feature
+0.436	9752_836
+0.386	9752_34914
+0.227	7755_9519
+0.206	9752_34914 9752_836
+0.184	7755_9519 7755_35295
+0.178	7738_2112
+0.174	9752_836 9752_34914
+0.174	7738_6486
+0.168	7755_35295
+0.166	7755_35295 7755_9519
… 286 more positive …
… 49705 more negative …

Weight^?	Feature
+0.524	9716_61258
+0.373	7716_17211
+0.279	7716_32974
+0.250	9716_61258 7716_17211
+0.219	7716_32952
+0.208	7716_32974 9716_61258
+0.168	7752_8571
+0.138	7716_17211 9716_61258
+0.133	9716_61258 7716_32974
+0.125	7765_33462
… 364 more positive …
… 49627 more negative …

Weight^?	Feature
+0.504	9716_2265
+0.436	9701_7835
+0.401	9716_2265 9716_2265
+0.349	9701_7835 9701_7835
+0.224	9716_2265 9716_2265 9716_2265
+0.168	9752_15576
+0.156	9701_27526
+0.155	9752_15570
+0.150	9764_21896
+0.145	9007_55646
… 19 more positive …
… 49972 more negative …

Weight^?	Feature
+0.574	7755_6972
+0.278	7755_6970
+0.216	7716_62504
+0.203	7716_33248
+0.193	9716_30646
+0.175	7772_775
+0.172	7755_6970 7755_6972
+0.159	7716_21029
+0.136	7716_2460
+0.132	9716_30646 7716_62504
… 204 more positive …
… 49787 more negative …

Weight^?	Feature
+0.302	9716_31062
+0.277	9752_14642
+0.272	9752_14641
+0.241	9752_3227
+0.226	9752_14641 9752_14642
+0.204	9700_30033
+0.189	9700_30159
+0.175	7730_380 9716_31062
+0.157	7730_3507
+0.155	9716_33838
… 110 more positive …
… 49881 more negative …

Weight^?	Feature
+0.446	9700_6532
+0.329	7700_4206
+0.267	7745_46879 7745_46879
+0.267	7745_46876
+0.262	7745_46879
+0.245	5061_1930 5061_1930
+0.201	7745_46879 7745_46879 7745_46879
+0.195	5061_1930
+0.184	5061_416
+0.172	7752_180
… 49 more positive …
… 49942 more negative …

Weight^?	Feature
+0.503	9700_8631
+0.397	7700_8356
+0.367	9700_8631 7700_8356
+0.276	7700_8356 9700_8631
+0.207	7762_32120
+0.184	7700_8356 9700_8631 7700_8356
+0.184	7762_32125
+0.184	9700_8631 7700_8356 9700_8631
+0.165	7784_18225
+0.148	7700_49383
… 118 more positive …
… 49873 more negative …

Weight^?	Feature
+0.443	7752_3249
+0.428	9752_3227
+0.320	9752_3227 7752_3249
+0.302	7746_33265
+0.250	9746_35024
+0.231	7752_3249 9752_3227
+0.217	7752_3250
+0.178	7746_33265 9746_35024
+0.160	9746_35024 7746_33265
+0.124	7752_3249 9752_3227 7752_3249
… 83 more positive …
… 49908 more negative …

Weight^?	Feature
+0.393	9739_41987
+0.272	9739_9217
+0.269	9739_48386
+0.269	9739_27909
+0.240	9935_53505
+0.237	9935_53516
+0.194	9935_8449
+0.158	7752_2306
+0.125	7755_2984
+0.120	9935_53505 9935_53516
… 905 more positive …
… 49086 more negative …

Weight^?	Feature
+0.305	9701_42242
+0.281	5008_15876
+0.262	9734_30466
+0.248	9734_30486
+0.206	5005_2927
+0.195	7747_6288
+0.189	9734_30486 9734_30466
+0.182	9701_42245
+0.179	9701_46339
+0.169	7700_1152
… 483 more positive …
… 49508 more negative …

Weight^?	Feature
+0.511	9739_61700
+0.360	7755_38450
+0.278	9739_61700 7755_38450
+0.256	7755_38450 9739_61700
+0.196	9739_61700 9739_61700
+0.191	9740_28676
+0.185	9739_61700 7755_38450 9739_61700
+0.174	7715_15371
+0.167	9937_62724
+0.153	5025_65284
… 485 more positive …
… 49506 more negative …

Weight^?	Feature
+0.397	9716_1797
+0.311	9716_33026
+0.309	9916_65284
+0.252	9716_1794
+0.247	9716_30212
+0.245	9916_65322
+0.181	9916_65322 9916_65284
+0.149	9716_30209
+0.149	7752_8747
+0.144	9716_33026 9716_1797
… 889 more positive …
… 49102 more negative …

Weight^?	Feature
+0.534	7757_2258
+0.360	9937_38145
+0.334	9937_38148
+0.294	7757_1154
+0.267	9937_55810
+0.201	9757_1144
+0.160	9937_38146
+0.153	9937_38156
+0.131	9757_1146
+0.131	9937_38148 7757_2258
… 318 more positive …
… 49673 more negative …

Weight^?	Feature
+0.473	5085_27553
+0.397	7754_26705
+0.279	9700_14256
+0.265	5085_27556
+0.196	9700_14256 7754_26705
+0.164	7754_26705 9700_14256
+0.160	9754_10283
+0.159	5012_17137
+0.138	7754_26703
+0.138	5085_27553 5085_27553
… 316 more positive …
… 49675 more negative …

Weight^?	Feature
+0.483	9935_22560
+0.428	9935_22531
+0.381	9935_22560 9935_22531
+0.368	9935_22531 9935_22560
+0.248	9935_22531 9935_22560 9935_22531
+0.240	9935_22560 9935_22531 9935_22560
+0.163	7716_17265
+0.109	9935_22560 9935_22560
+0.100	9935_22531 9935_22531
+0.099	9716_61956
… 556 more positive …
… 49435 more negative …

Weight^?	Feature
+0.419	9937_37644
+0.379	9937_37633
+0.250	9937_37635
+0.244	9937_37633 9937_37644
+0.229	9739_24326
+0.221	9739_9217
+0.209	9937_37644 9937_37633
+0.158	5074_64629
+0.144	9739_24326 9739_9217
+0.128	9739_9217 9739_24326
… 656 more positive …
… 49335 more negative …

Weight^?	Feature
+0.416	5015_56952
+0.326	9752_15576
+0.290	7755_35295
+0.265	9752_15570
+0.223	5015_53391
+0.205	7755_9519
+0.195	5069_53397
+0.149	5015_56952 5069_53397
+0.149	5069_53392
+0.149	7752_15573
… 187 more positive …
… 49804 more negative …

Weight^?	Feature
+0.454	9797_12971
+0.382	7797_46515
+0.302	7797_46515 9797_12971
+0.291	9797_12971 7797_46515
+0.199	7797_46515 9797_12971 7797_46515
+0.168	9797_12971 7797_46515 9797_12971
+0.168	9797_12971 9797_12971
+0.154	7743_33033
+0.139	7797_46515 7797_46515 9797_12971
+0.132	7716_745
… 222 more positive …
… 49769 more negative …

Weight^?	Feature
+0.457	9716_37506
+0.448	9716_37501
+0.237	9716_37506 9716_37501
+0.220	9716_1841
+0.206	7716_2147
+0.187	7716_2147 9716_37506
+0.178	9742_28347
+0.174	9752_47576
+0.145	9716_37506 7716_2147
+0.145	9716_1841 9716_37501
… 164 more positive …
… 49827 more negative …

Weight^?	Feature
+0.347	7752_28340
+0.336	7742_39295
+0.328	7716_44972
+0.250	7716_44971
+0.250	7700_33799
+0.223	7746_14898
+0.194	7742_39295 7752_28340
+0.164	7752_28340 7742_39295
+0.160	7716_44972 7716_44972
+0.154	7700_4700
… 233 more positive …
… 49758 more negative …

Weight^?	Feature
+0.466	9716_62468
+0.398	9716_13059
+0.227	5031_24577
+0.211	9934_39425
+0.210	9716_25376
+0.186	9716_25347
+0.181	9716_62468 9716_13059
+0.179	9716_13059 9716_62468
+0.178	5011_12377
+0.134	9716_28673
… 655 more positive …
… 49336 more negative …

Weight^?	Feature
+0.373	7752_46005
+0.317	7716_1840
+0.243	7716_25988
+0.232	9716_6429
+0.228	9716_33844
+0.197	9752_10360
+0.193	7716_1843
+0.186	7752_10367
+0.183	7742_54601
+0.165	7752_14305
… 168 more positive …
… 49823 more negative …

Weight^?	Feature
+0.266	7755_38818
+0.246	5061_1581
+0.236	5061_1930
+0.229	5061_1588
+0.222	5061_1790
+0.206	5061_1584 5061_1588
+0.195	5061_1587
+0.177	5061_1692
+0.174	5061_1584
+0.171	5061_1693
… 121 more positive …
… 49870 more negative …

Weight^?	Feature
+0.654	7716_6434
+0.311	9716_39387
+0.299	9716_4101
+0.211	7716_6434 7716_6434
+0.159	7716_6434 9716_39387
+0.156	9716_4101 7716_6434
+0.151	7716_6434 9716_4101
+0.147	9716_39387 7716_6434
+0.143	9716_19972
+0.128	7716_3976
… 623 more positive …
… 49368 more negative …

Weight^?	Feature
+0.363	9909_39684
+0.355	9909_64769
+0.291	9935_48385
+0.285	9916_51457
+0.256	9909_39684 9909_64769
+0.236	9909_64769 9909_39684
+0.207	9716_33832
+0.149	9909_39684 9909_64769 9909_39684
+0.149	9909_12034
+0.147	9716_7172
… 689 more positive …
… 49302 more negative …

Weight^?	Feature
+0.341	9732_40196
+0.305	9916_9985
+0.271	7700_33791
+0.266	9732_40193
+0.256	9701_49156
+0.249	9916_9996
+0.205	9701_49153
+0.191	7700_33793
+0.151	9916_9987
+0.128	7755_14939
… 812 more positive …
… 49179 more negative …

Weight^?	Feature
+0.383	7743_2582
+0.281	9716_6430
+0.255	9716_1822
+0.242	9752_23293
+0.239	7716_6440
+0.204	9743_2546
+0.201	9752_980
+0.189	5048_56775
+0.189	9716_1822 9716_6430
+0.182	9743_2544
… 285 more positive …
… 49706 more negative …

Weight^?	Feature
+0.445	5033_3467
+0.333	9033_40874
+0.320	9752_509
+0.259	9033_40874 5033_3467
+0.241	7752_517
+0.191	9746_37926
+0.185	9752_509 7752_517
+0.185	5033_3467 9033_40874
+0.174	7746_31256
+0.165	9746_38504
… 175 more positive …
… 49816 more negative …

Weight^?	Feature
+0.416	7716_3669
+0.384	7755_307
+0.305	7755_4963
+0.278	7755_307 7716_3669
+0.195	7716_3669 7755_307
+0.185	7755_57579
+0.178	7755_55451
+0.174	7752_45684
+0.158	7752_8571
+0.143	7752_846
… 226 more positive …
… 49765 more negative …

Weight^?	Feature
+0.387	9794_58114
+0.365	9794_60674
+0.260	9734_30486
+0.232	9794_60694
+0.227	9734_30466 9734_30486
+0.226	9734_30486 9734_30466
+0.210	9734_30466
+0.203	9916_59905
+0.186	9734_30466 9734_30486 9734_30466
+0.170	9734_30486 9734_30466 9734_30486
… 489 more positive …
… 49502 more negative …

Weight^?	Feature
+0.465	9734_14880
+0.456	9734_14851
+0.367	9734_14880 9734_14851
+0.367	9734_14851 9734_14880
+0.283	9734_14880 9734_14851 9734_14880
+0.250	9734_14851 9734_14880 9734_14851
+0.151	9734_14880 9734_14880
+0.119	9734_14880 9734_14880 9734_14851
+0.117	9734_14851 9734_14880 9734_14880
+0.110	9734_14851 9734_14851
… 621 more positive …
… 49370 more negative …

Weight^?	Feature
+0.493	7716_39211
+0.394	9734_39958
+0.303	9701_45313
+0.196	9734_39938
+0.158	9716_65299
+0.158	9716_65298
+0.154	9701_24854
+0.154	9701_24837
+0.143	9701_24834
+0.123	7716_39211 9734_39958
… 631 more positive …
… 49360 more negative …

Weight^?	Feature
+0.385	9937_46604
+0.358	9937_46593
+0.271	9937_46604 9937_46593
+0.271	9937_46593 9937_46604
+0.226	9734_33284
+0.215	9734_33286
+0.206	9937_46604 9937_46593 9937_46604
+0.206	9937_46593 9937_46604 9937_46593
+0.171	9734_12801
+0.147	9734_12812
… 812 more positive …
… 49179 more negative …

Weight^?	Feature
+0.426	9937_43011
+0.385	9937_43040
+0.328	9937_43040 9937_43011
+0.319	9937_43011 9937_43040
+0.263	9937_43011 9937_43040 9937_43011
+0.250	9937_43040 9937_43011 9937_43040
+0.185	9716_39941
+0.175	9739_22531
+0.173	9937_34817
+0.111	9716_39941 9739_22531
… 887 more positive …
… 49104 more negative …

Weight^?	Feature
+0.456	9742_46003
+0.355	7755_19649
+0.353	9742_21276
+0.274	9742_46003 9742_21276
+0.209	9742_21276 9742_46003
+0.196	7755_19994
+0.183	9742_25658
+0.171	9742_25657
+0.157	9742_22555 9742_46003
+0.157	9742_46003 9742_21276 9742_46003
… 228 more positive …
… 49763 more negative …

Weight^?	Feature
+0.432	9734_5378
+0.382	9734_5378 9734_5398
+0.365	9734_5398 9734_5378
+0.349	9734_5398
+0.300	9734_5398 9734_5378 9734_5398
+0.294	9734_5378 9734_5398 9734_5378
+0.124	9734_12801
+0.111	7716_3252
+0.102	9734_33283
+0.088	9734_12801 9734_5378
… 935 more positive …
… 49056 more negative …

Weight^?	Feature
+0.518	9716_26295
+0.322	7755_33269
+0.298	7755_33269 9716_26295
+0.284	9716_26295 7755_33269
+0.258	9953_37016
+0.203	9716_31989
+0.176	7755_33269 9716_26295 7755_33269
+0.157	9953_37016 9716_26295
+0.149	9716_26295 7755_33269 9716_26295
+0.142	9716_26295 9953_37016
… 184 more positive …
… 49807 more negative …

Weight^?	Feature
+0.353	9716_15875
+0.323	9736_48386
+0.312	9736_48406
+0.298	9716_15904
+0.240	9716_50689
+0.238	9736_48406 9736_48386
+0.234	9736_48386 9736_48406
+0.171	9736_27137
+0.145	9716_15875 9716_15904
+0.141	9716_15875 9716_50689
… 758 more positive …
… 49233 more negative …

Weight^?	Feature
+0.346	5061_10365
+0.261	5061_1731
+0.258	5061_1777
+0.224	7772_1255
+0.217	5061_960
+0.194	7772_101
+0.169	5061_1631
+0.169	5061_1779
+0.166	7730_2745
+0.160	5061_1721
… 119 more positive …
… 49872 more negative …

Weight^?	Feature
+0.610	7716_60475
+0.281	9716_32507
+0.276	7716_60475 7716_60475
+0.268	9716_31427
+0.245	9716_31427 7716_60475
+0.243	7716_60475 9716_31427
+0.201	9716_32507 7716_60475
+0.201	7716_60475 9716_32507
+0.140	7716_60475 9716_31427 7716_60475
+0.129	7716_60475 7716_60475 9716_31427
… 174 more positive …
… 49817 more negative …

Weight^?	Feature
+0.477	9716_17232
+0.319	5018_60001
+0.319	5022_51696
+0.265	9716_17232 9716_17232
+0.247	9752_870
+0.192	5018_50832
+0.180	9754_12579
+0.167	9716_31063
+0.128	5022_9923
+0.128	9754_46662
… 131 more positive …
… 49860 more negative …

Weight^?	Feature
+0.554	7701_4998
+0.484	7701_4998 7701_4998
+0.415	7701_4998 7701_4998 7701_4998
+0.245	7716_32952
+0.195	7787_44160
+0.119	7799_2566
+0.098	7716_61255
+0.086	9716_61253
+0.079	7752_6246
+0.071	7730_460
… 50 more positive …
… 49941 more negative …

Weight^?	Feature
+0.562	7754_2891
+0.366	7700_801
+0.296	9701_2817
+0.176	9716_1797
+0.173	9016_50311
+0.166	9754_1671
+0.155	5016_5566
+0.143	9701_40451
+0.142	7754_2891 7700_801
+0.127	7754_2891 7754_2891
… 557 more positive …
… 49434 more negative …

Weight^?	Feature
+0.378	9916_42753
+0.331	9916_42764
+0.307	5047_12545
+0.292	9916_42753 9916_42764
+0.268	9916_42764 9916_42753
+0.242	9916_42753 9916_42764 9916_42753
+0.231	9916_42764 9916_42753 9916_42764
+0.193	5047_45313
+0.186	5032_7579
+0.164	9031_5132
… 630 more positive …
… 49361 more negative …

Weight^?	Feature
+0.399	9716_3841
+0.292	9716_50689
+0.289	9937_32260
+0.286	9790_59393
+0.240	9716_3841 9716_50689
+0.228	9716_50689 9716_3841
+0.195	9790_38916
+0.187	7716_49282
+0.159	9716_4102
+0.152	9942_15873
… 861 more positive …
… 49130 more negative …

Weight^?	Feature
+0.623	9935_14850
+0.357	9752_10502
+0.319	9752_10502 9935_14850
+0.310	9935_14850 9752_10502
+0.220	9935_14850 9935_14850
+0.195	9935_14850 9752_10502 9935_14850
+0.163	9752_10502 9935_14850 9752_10502
+0.131	9716_55558
+0.127	9752_10507
+0.120	9935_14850 9935_14850 9752_10502
… 363 more positive …
… 49628 more negative …

Weight^?	Feature
+0.444	7752_7903
+0.381	9934_8963
+0.321	9916_1794
+0.312	9916_1796
+0.210	9916_1814
+0.192	9734_16641
+0.180	7752_49241
+0.130	7752_7903 9934_8963
+0.126	9934_8961
+0.126	7752_7904
… 838 more positive …
… 49153 more negative …

Weight^?	Feature
+0.339	9900_64001
+0.318	9900_64012
+0.263	9716_33093
+0.263	9734_48386
+0.260	9734_48406
+0.212	9900_64012 9900_64001
+0.198	9900_64001 9900_64012
+0.146	9734_48388
+0.131	9900_64012 9900_64001 9900_64012
+0.130	9734_30467
… 714 more positive …
… 49277 more negative …

Weight^?	Feature
+0.361	9739_52741
+0.332	9934_32257
+0.318	9739_52738
+0.260	9739_52758
+0.172	9739_52741 9739_52738
+0.172	9916_20993
+0.155	5010_54991
+0.146	9739_52738 9739_52741
+0.139	9916_55809
+0.126	9074_48900
… 922 more positive …
… 49069 more negative …

Weight^?	Feature
+0.513	9916_20226
+0.460	9916_20246
+0.345	9916_20246 9916_20226
+0.340	9916_20226 9916_20246
+0.252	9916_20226 9916_20246 9916_20226
+0.241	9916_20246 9916_20226 9916_20246
+0.097	9916_15884
+0.097	9916_20226 9916_20226
+0.092	9916_20246 9916_20246
+0.079	9716_28684
… 660 more positive …
… 49331 more negative …

Weight^?	Feature
+0.503	9935_63776
+0.468	9935_63747
+0.337	9935_63776 9935_63747
+0.336	9935_63747 9935_63776
+0.218	9935_63776 9935_63747 9935_63776
+0.205	9935_63747 9935_63776 9935_63747
+0.166	9935_63750
+0.102	9935_63776 9935_63776
+0.099	9935_59916
+0.088	9734_5388
… 731 more positive …
… 49260 more negative …

Weight^?	Feature
+0.370	7716_55676
+0.346	7716_31789
+0.333	7757_2653
+0.272	9953_6544
+0.211	7746_6550
+0.204	9746_16330
+0.178	5022_19797
+0.175	9757_2637
+0.159	9953_6544 7716_55676
+0.158	5022_19797 5022_19797
… 133 more positive …
… 49858 more negative …

Weight^?	Feature
+0.336	5025_57348
+0.261	9776_35331
+0.259	7752_762
+0.207	9776_35360
+0.200	9916_52225
+0.192	9041_278
+0.162	7752_6247
+0.161	9734_33312
+0.159	9041_37142
+0.149	9916_52229
… 742 more positive …
… 49249 more negative …

Weight^?	Feature
+0.406	9937_55328
+0.396	9937_55299
+0.376	9937_55299 9937_55328
+0.373	9937_55328 9937_55299
+0.304	9937_55328 9937_55299 9937_55328
+0.279	9937_55299 9937_55328 9937_55299
+0.245	9732_63233
+0.135	9732_61187
+0.094	9732_61187 9732_63233
+0.091	9937_55299 9937_55299
… 571 more positive …
… 49420 more negative …

Weight^?	Feature
+0.534	5018_19769
+0.321	7752_1389
+0.229	9916_15876
+0.191	5018_19766
+0.190	5018_19762
+0.179	9011_35074
+0.155	7759_30123
+0.143	9047_42303
+0.137	9716_7180
+0.136	7752_1388
… 167 more positive …
… 49824 more negative …

Weight^?	Feature
+0.626	7755_33346
+0.258	9752_14633
+0.235	9700_8607
+0.219	7755_33346 7755_33346
+0.209	7755_38846 7755_33346
+0.193	7755_38846
+0.136	9700_13319
+0.134	7755_33346 9700_30157
+0.132	7755_35296
+0.131	9752_14633 9700_8607
… 121 more positive …
… 49870 more negative …

Weight^?	Feature
+0.429	7755_35295
+0.422	7755_14939
+0.286	7755_33344
+0.254	7755_14939 7755_33344
+0.239	7755_14939 7755_35295
+0.228	7755_35295 7755_14939
+0.227	7755_35888
+0.148	7755_33346
+0.141	7755_35295 7755_14939 7755_33344
+0.141	7755_35888 7755_14939
… 155 more positive …
… 49836 more negative …

Weight^?	Feature
+0.419	9716_33842
+0.302	9716_33847
+0.279	9716_33842 9716_33842
+0.253	9716_33842 9716_33842 9716_33842
+0.242	9716_6428
+0.226	9752_12947
+0.207	7730_3803
+0.202	9752_14633 9752_12947
+0.180	7730_3803 7730_3507
+0.161	7755_15802
… 88 more positive …
… 49903 more negative …

Weight^?	Feature
+0.456	7755_861
+0.433	7752_3251
+0.232	7755_36181
+0.223	7755_863
+0.206	7752_3251 9752_2903
+0.180	7760_36465
+0.178	7752_5130
+0.166	9752_2903
+0.165	7760_56271
+0.159	7755_865
… 144 more positive …
… 49847 more negative …

Weight^?	Feature
+0.383	9746_31872
+0.315	7746_1772
+0.260	9758_17184
+0.207	9716_19990
+0.206	9758_28166
+0.162	9902_52748
+0.162	9758_28192
+0.162	9758_17155
+0.161	9716_19970
+0.145	9754_12576
… 807 more positive …
… 49184 more negative …

Weight^?	Feature
+0.371	9734_12801
+0.314	9734_12812
+0.258	9734_12812 9734_12801
+0.253	9734_12801 9734_12812
+0.231	9734_12812 9734_12801 9734_12812
+0.226	9916_65026
+0.220	9916_65046
+0.220	9734_12801 9734_12812 9734_12801
+0.159	9734_33281
+0.144	9916_65046 9916_65026
… 925 more positive …
… 49066 more negative …

Weight^?	Feature
+0.474	9935_44292
+0.459	7752_2306
+0.301	7752_2306 9935_44292
+0.282	9935_44292 7752_2306
+0.254	9935_8449
+0.164	7752_2306 9935_44292 7752_2306
+0.156	9935_44292 7752_2306 9935_44292
+0.132	9732_39172
+0.122	9935_8449 9935_44292
+0.113	9935_44292 9935_8449
… 368 more positive …
… 49623 more negative …

Weight^?	Feature
+0.473	9934_56321
+0.361	9934_35846
+0.334	9934_35843
+0.269	7752_3622
+0.235	9934_52225
+0.212	9934_52236
+0.167	9934_35843 9934_35846
+0.161	9934_35846 9934_35843
+0.157	9934_35872
+0.143	7752_36201
… 637 more positive …
… 49354 more negative …

Weight^?	Feature
+0.650	9005_44856
+0.325	9026_41063
+0.325	9026_49321
+0.217	9026_49321 9005_44856
+0.217	9026_41063 9005_44856
+0.208	9005_44856 9026_49321
+0.192	9005_44856 9026_41063
+0.150	9005_44856 9026_49321 9005_44856
+0.108	9005_44856 9026_41063 9005_44856
+0.108	9005_44856 9005_44856
… 76 more positive …
… 49915 more negative …

Weight^?	Feature
+0.374	9909_50689
+0.356	9909_50700
+0.289	9937_59404
+0.246	9937_59393
+0.192	9909_50689 9909_50700
+0.184	9937_59393 9937_59404
+0.174	9909_50700 9909_50689
+0.172	9937_59404 9937_59393
+0.166	9734_37633
+0.145	9734_61702
… 785 more positive …
… 49206 more negative …

Weight^?	Feature
+0.378	9732_27138
+0.370	9732_27138 9732_27158
+0.368	9732_27158 9732_27138
+0.348	9732_27158
+0.293	9732_27158 9732_27138 9732_27158
+0.293	9732_27138 9732_27158 9732_27138
+0.173	9754_43521
+0.173	9732_27141
+0.169	9754_43532
+0.121	9754_43532 9754_43521
… 665 more positive …
… 49326 more negative …

Weight^?	Feature
+0.392	7700_6190
+0.370	9701_64772
+0.275	7700_6191
+0.202	9701_24864
+0.184	9701_14082
+0.179	9701_45314
+0.177	9701_14102
+0.177	9900_26625
+0.159	9701_24835
+0.154	9701_64780
… 578 more positive …
… 49413 more negative …

Weight^?	Feature
+0.566	7700_5871
+0.390	7700_55154
+0.216	9934_27137
+0.212	9732_27137
+0.160	9732_27148
+0.148	9934_27137 5050_16641
+0.147	9934_29445
+0.146	7700_55154 7700_5871
+0.136	7700_5871 7700_55154
+0.136	5050_16641
… 379 more positive …
… 49612 more negative …

Weight^?	Feature
+0.348	9716_1797
+0.308	9716_43778
+0.185	9913_45846
+0.179	9701_42242
+0.159	9913_45826
+0.158	9716_1796
+0.147	9716_64516
+0.136	9735_13059
+0.134	9735_13088
+0.130	9716_43778 9716_1797
… 1012 more positive …
… 48979 more negative …

Weight^?	Feature
+0.539	9935_53505
+0.519	9935_8449
+0.325	9935_8449 9935_53505
+0.316	9935_53505 9935_8449
+0.183	9935_53505 9935_8449 9935_53505
+0.182	9935_8449 9935_53505 9935_8449
+0.143	9935_8449 9935_8449
+0.133	9935_53516
+0.128	9935_53505 9935_53505
+0.109	9935_8449 9935_53505 9935_53505
… 512 more positive …
… 49479 more negative …

Weight^?	Feature
+0.607	7746_31431
+0.337	7716_65294
+0.256	7716_902
+0.255	7716_2268
+0.224	7746_31431 7746_31431
+0.184	7716_65294 7746_31431
+0.181	7746_31431 7716_65294
+0.178	7746_31431 7716_902
+0.173	7746_31431 7716_2268
+0.165	7716_2268 7746_31431
… 327 more positive …
… 49664 more negative …

Weight^?	Feature
+0.511	7716_44974
+0.252	5091_64313
+0.242	9091_52009
+0.227	9716_33092
+0.222	9716_33092 7716_44974
+0.199	7716_44974 9716_33092
+0.186	9716_33093
+0.163	9091_52009 5091_64313
+0.160	7752_21238
+0.158	5091_64313 9091_52009
… 452 more positive …
… 49539 more negative …

Weight^?	Feature
+0.495	9716_3151
+0.490	7716_3161
+0.268	9716_3151 7716_3161
+0.259	7716_3161 9716_3151
+0.220	7716_3161 7716_3161
+0.171	9716_3151 9716_3151
+0.153	9934_56076
+0.141	7716_3161 9716_3151 7716_3161
+0.126	9716_3151 7716_3161 9716_3151
+0.126	9716_3151 7716_3161 7716_3161
… 601 more positive …
… 49390 more negative …

Weight^?	Feature
+0.594	5024_42692
+0.490	9024_63888
+0.200	5024_42692 5024_42692
+0.190	9024_63888 5024_42692
+0.179	7752_15944
+0.175	5024_42692 9024_63888
+0.172	9024_63891
+0.172	5024_7820
+0.154	9024_42697
+0.147	9024_63888 9024_63888
… 448 more positive …
… 49543 more negative …

Weight^?	Feature
+0.397	5061_1930
+0.289	7760_46585
+0.286	5061_1693
+0.222	5061_1690
+0.186	5061_414
+0.186	5061_1930 5061_1930
+0.184	7755_30220
+0.180	5061_1187
+0.166	5061_1928
+0.163	5061_1171
… 73 more positive …
… 49918 more negative …

Weight^?	Feature
+0.580	9716_31987
+0.346	9716_31987 9716_31987
+0.252	9752_31489
+0.230	9752_14633 9752_31489
+0.201	7755_38846 9716_33847
+0.180	9752_31489 9700_8601
+0.168	9716_31987 9716_31987 7755_15806
+0.168	9716_31986 9716_31987
+0.168	9716_31987 7755_15806
+0.159	9716_33847
… 100 more positive …
… 49891 more negative …

Weight^?	Feature
+0.447	7753_62669
+0.409	7716_3579
+0.343	7753_6136
+0.312	7716_3579 7753_62669
+0.306	7753_62669 7716_3579
+0.181	7716_3579 7753_62669 7716_3579
+0.152	7753_62669 7716_3579 7753_62669
+0.143	7716_3579 7753_6136
+0.138	7753_6136 7716_3579
+0.125	7716_3579 7716_3579
… 358 more positive …
… 49633 more negative …

Weight^?	Feature
+0.650	7755_306
+0.321	7755_37277
+0.243	7755_37277 7755_306
+0.225	7755_306 7755_37277
+0.224	7755_306 7755_306
+0.215	7755_32043
+0.189	7755_32043 7755_306
+0.150	7755_26527
+0.140	7755_306 7755_32043
+0.125	7755_306 7755_37277 7755_306
… 168 more positive …
… 49823 more negative …

Weight^?	Feature
+0.424	9739_24322
+0.377	9739_23809
+0.355	9739_24325
+0.320	9739_24342
+0.226	9739_24322 9739_24342
+0.219	9739_24342 9739_24322
+0.158	9739_24322 9739_24342 9739_24322
+0.148	9739_24325 9739_23809
+0.143	9739_23809 9739_24325
+0.133	9916_55809
… 814 more positive …
… 49177 more negative …

Weight^?	Feature
+0.383	9937_19717
+0.346	9937_19714
+0.346	9937_40194
+0.283	9716_19974
+0.243	9937_19734
+0.168	9937_40194 9937_19717
+0.168	9937_19717 9937_40194
+0.135	9736_55052
+0.133	9937_19717 9937_19714
+0.133	9716_19973
… 756 more positive …
… 49235 more negative …

Weight^?	Feature
+0.381	9916_2305
+0.374	9916_2316
+0.280	9716_7172
+0.262	9916_2305 9916_2316
+0.254	9916_2316 9916_2305
+0.196	9716_7173 9716_7172
+0.194	9916_2316 9916_2305 9916_2316
+0.194	9916_2305 9916_2316 9916_2305
+0.194	9716_7172 9716_7173
+0.168	9716_7172 9716_7173 9716_7172
… 900 more positive …
… 49091 more negative …

Weight^?	Feature
+0.419	9739_24324
+0.378	9739_41985
+0.234	9739_24332
+0.205	9916_20236
+0.198	9916_20225
+0.197	9739_41985 9739_24324
+0.178	9739_24321
+0.168	9739_24324 9739_41985
+0.165	9916_15875
+0.159	9916_65292
… 947 more positive …
… 49044 more negative …

Weight^?	Feature
+0.503	9734_61702
+0.351	9734_61699
+0.315	9734_61728
+0.300	9734_59905
+0.199	9734_59905 9734_61702
+0.188	9734_61702 9734_59905
+0.165	9734_61699 9734_61728
+0.131	9734_61702 9734_61699
+0.131	9734_61728 9734_61699
+0.120	9734_61702 9734_61728
… 573 more positive …
… 49418 more negative …

Weight^?	Feature
+0.336	9734_30486
+0.322	9734_30466 9734_30486
+0.312	9734_30486 9734_30466
+0.267	9734_30466
+0.260	9734_30486 9734_30466 9734_30486
+0.259	9734_30466 9734_30486 9734_30466
+0.245	7752_766
+0.215	9916_52227
+0.187	9916_52246
+0.183	9735_54533
… 430 more positive …
… 49561 more negative …

Weight^?	Feature
+0.537	9916_59680
+0.511	9916_59651
+0.320	9916_59680 9916_59651
+0.290	9916_59651 9916_59680
+0.200	9916_59680 9916_59651 9916_59680
+0.196	9916_59651 9916_59680 9916_59651
+0.108	9916_3350
+0.108	9916_3330
+0.102	9916_59680 9916_59680
+0.098	9916_32769
… 722 more positive …
… 49269 more negative …

Weight^?	Feature
+0.271	9935_63748
+0.264	9733_59396
+0.229	9043_53510
+0.229	5064_36357
+0.181	9735_25603
+0.165	5046_54149
+0.160	9935_14336
+0.157	9733_23971
+0.151	7752_61575
+0.150	9736_11782
… 776 more positive …
… 49215 more negative …

Weight^?	Feature
+0.334	9732_27142
+0.283	7755_872
+0.239	9734_14337
+0.192	9732_12289
+0.172	7755_873
+0.171	9734_34817
+0.167	9734_30465
+0.134	9734_30465 9734_61702 9734_30465
+0.133	9734_14340
+0.132	9734_61702 9734_30465
… 899 more positive …
… 49092 more negative …

Weight^?	Feature
+0.320	9739_29953
+0.319	9754_40450
+0.312	9916_46339
+0.291	9754_40470
+0.290	7755_38450
+0.254	9754_40452
+0.212	9916_20996
+0.171	9916_20996 9916_46339
+0.158	9916_46339 9916_20996
+0.151	5046_64340
… 608 more positive …
… 49383 more negative …

Weight^?	Feature
+0.604	5069_61443
+0.447	5015_51628
+0.314	5015_61397
+0.240	5015_51628 5069_61443
+0.236	5069_61443 5015_51628
+0.170	5069_61443 5015_61397
+0.165	5015_61397 5069_61443
+0.161	5069_61443 5069_61443
+0.128	5015_51628 5015_51628
+0.108	5069_61448
… 263 more positive …
… 49728 more negative …

Weight^?	Feature
+0.374	5047_8546
+0.321	7757_7040
+0.258	7755_9520
+0.239	7757_7033
+0.228	7755_44696
+0.205	7757_3276
+0.198	7755_35789
+0.196	9757_1050
+0.174	7755_57579
+0.150	9752_1936
… 115 more positive …
… 49876 more negative …

Weight^?	Feature
+0.563	7716_3161
+0.309	9716_3151
+0.283	9734_5382
+0.236	9916_42754
+0.221	9916_42774
+0.198	7716_3161 9716_3151
+0.195	9716_3151 7716_3161
+0.179	9716_4102
+0.156	9916_42774 9916_42754
+0.152	9916_42754 9916_42774
… 739 more positive …
… 49252 more negative …

Weight^?	Feature
+0.402	7716_55676
+0.355	7746_6551
+0.342	7716_39675
+0.304	7746_6552
+0.269	7746_6550
+0.170	7716_28931
+0.169	7716_31789
+0.157	7716_55676 7746_6551
+0.156	7746_6552 7716_39675
+0.143	7716_39675 7746_6552
… 507 more positive …
… 49484 more negative …

Weight^?	Feature
+0.618	9741_48901
+0.363	7741_5928
+0.270	7712_45270
+0.196	7746_14896
+0.194	9741_48901 7741_5928
+0.192	7741_5926
+0.188	7741_5928 9741_48901
+0.152	9741_48901 9741_48901
+0.151	7712_45270 9741_48901
+0.138	9741_48901 7712_45270
… 415 more positive …
… 49576 more negative …

Weight^?	Feature
+0.403	9716_1796
+0.323	9716_64513
+0.281	9716_64513 9716_64524 9716_64513
+0.253	9716_1793
+0.250	9716_64524 9716_64513
+0.237	9716_64513 9716_64524
+0.236	9716_64524 9716_64513 9716_64524
+0.233	9716_64524
+0.164	9716_64516
+0.157	9937_59393
… 386 more positive …
… 49605 more negative …

Weight^?	Feature
+0.340	9734_60694
+0.326	9916_63747
+0.322	9916_63776
+0.313	9734_60674
+0.272	9916_63776 9916_63747
+0.259	9916_63747 9916_63776
+0.230	9734_60694 9734_60674
+0.230	9734_60677
+0.218	9916_63776 9916_63747 9916_63776
+0.214	9916_63747 9916_63776 9916_63747
… 786 more positive …
… 49205 more negative …

Weight^?	Feature
+0.462	9790_38913
+0.408	9790_38924
+0.244	9790_38924 9790_38913
+0.238	7752_61571
+0.229	9790_38913 9790_38924
+0.209	9790_38916
+0.181	9754_18690
+0.180	9735_7169
+0.159	9935_63748
+0.135	9752_13216
… 678 more positive …
… 49313 more negative …

Weight^?	Feature
+0.643	9716_2260 9716_2260
+0.599	9716_2260 9716_2260 9716_2260
+0.463	9716_2260
… 1 more positive …
… 49990 more negative …
-0.006	9702_30576
-0.006	9700_8607
-0.006	9916_20226
-0.006	9734_61702
-0.007	9702_30571
-0.008	9752_14642
-0.009	9752_14633

Weight^?	Feature
+0.717	7716_3158
+0.207	7716_3157
+0.199	7716_3158 7716_3157
+0.199	7716_3157 7716_3158
+0.177	9716_33107
+0.168	9716_3146
+0.146	7752_45995
+0.146	7716_3976
+0.134	7716_3158 9716_3146
+0.128	7716_6440
… 154 more positive …
… 49837 more negative …

Weight^?	Feature
+0.484	9739_34305
+0.341	9754_51203
+0.314	9754_51203 9739_34305
+0.271	9739_34305 9754_51203
+0.174	9760_27649
+0.170	9760_15110
+0.163	9739_34305 9754_51203 9739_34305
+0.155	9754_51232
+0.150	9934_22273
+0.149	9934_49665
… 470 more positive …
… 49521 more negative …

Weight^?	Feature
+0.513	9716_62011
+0.452	7716_31050
+0.335	9716_33828
+0.211	7752_7698
+0.201	7716_55654
+0.193	7752_7699
+0.193	7716_31050 9716_62011
+0.145	9716_62011 9716_33828
+0.138	7716_55654 9716_62011
+0.131	9716_33828 7716_31050
… 86 more positive …
… 49905 more negative …

Weight^?	Feature
+0.421	9734_16642
+0.356	9734_61700
+0.305	9734_16642 9734_61700
+0.251	9734_61700 9734_16642
+0.230	5007_770
+0.199	9734_16642 9734_61700 9734_16642
+0.167	5007_11521
+0.146	9734_61700 9734_16642 9734_61700
+0.143	5007_14594
+0.138	7716_63228
… 700 more positive …
… 49291 more negative …

Weight^?	Feature
+0.632	7716_3668
+0.382	7716_3669
+0.265	7716_3668 7716_3669
+0.199	7716_3669 7716_3668
+0.158	7752_759
+0.139	7716_20068
+0.139	9716_3664
+0.139	7716_3668 7716_3668
+0.123	9716_3664 7716_3668
+0.121	7716_3669 7716_20068
… 190 more positive …
… 49801 more negative …

Weight^?	Feature
+0.375	7755_55462
+0.329	5091_55862
+0.307	5091_57624
+0.270	7755_33264
+0.213	7755_55462 7755_33264
+0.190	7755_33264 7755_55462
+0.182	9091_55807
+0.136	9091_57627
+0.136	5091_55862 5091_55862
+0.132	7755_55462 7755_55462
… 452 more positive …
… 49539 more negative …

Weight^?	Feature
+0.380	9916_42764
+0.339	9916_42753
+0.333	9916_42764 9916_42753
+0.320	9916_42764 9916_42753 9916_42764
+0.317	9916_42753 9916_42764
+0.258	9916_42753 9916_42764 9916_42753
+0.164	9734_48406
+0.156	9716_33092
+0.150	9716_10497
+0.125	9759_22017
… 274 more positive …
… 49717 more negative …

Weight^?	Feature
+0.409	7755_2994
+0.323	7716_21029
+0.259	9735_21280
+0.241	9735_21251
+0.180	9735_41731
+0.177	7755_2994 7755_2994
+0.166	7755_2994 7716_21029
+0.166	7716_21029 7755_2994
+0.165	7755_2998
+0.158	9739_27907
… 574 more positive …
… 49417 more negative …

Weight^?	Feature
+0.447	9935_14597
+0.446	9935_14614
+0.384	9935_11523
+0.319	9935_14594
+0.187	9935_11523 9935_14597
+0.177	9935_14614 9935_11523
+0.166	9935_14597 9935_11523
+0.156	9935_14597 9935_14614
+0.137	9935_11523 9935_14614
+0.129	9935_11523 9935_14594
… 425 more positive …
… 49566 more negative …

Weight^?	Feature
+0.353	7716_11855
+0.343	7716_11870
+0.331	7716_11855 7716_11870
+0.311	7716_11870 7716_11855
+0.289	7757_12699
+0.270	7757_12693
+0.232	7716_11855 7716_11870 7716_11855
+0.232	7716_11870 7716_11855 7716_11870
+0.193	7757_12693 7757_12699
+0.165	7757_12699 7757_12693
… 116 more positive …
… 49875 more negative …

Weight^?	Feature
+0.537	7716_33418
+0.391	9716_31426
+0.254	7716_4759
+0.246	7716_62504
+0.232	9716_30646
+0.213	7716_33418 9716_31426
+0.210	9716_31426 7716_33418
+0.174	7716_61907
+0.122	7716_33418 7716_4759
+0.121	7716_6975
… 313 more positive …
… 49678 more negative …

Weight^?	Feature
+0.389	9934_20512
+0.328	9934_20483
+0.298	9736_48396
+0.262	9734_3078
+0.192	9934_20512 9736_48396
+0.185	9734_3075
+0.178	9736_48385
+0.175	9734_3104
+0.167	9736_48396 9934_20512
+0.160	9736_11781
… 395 more positive …
… 49596 more negative …

Weight^?	Feature
+0.434	7755_6982
+0.380	7755_15800
+0.273	7716_3668
+0.266	7755_6984
+0.265	7755_6592
+0.209	7755_6982 7755_15800
+0.199	7755_31332
+0.170	7716_3667
+0.167	7755_6983
+0.158	7755_15800 7755_6982
… 169 more positive …
… 49822 more negative …

Weight^?	Feature
+0.365	9716_30209
+0.343	9716_30220
+0.326	9716_30209 9716_30220
+0.319	9716_30220 9716_30209
+0.304	9716_30220 9716_30209 9716_30220
+0.297	9716_30209 9716_30220 9716_30209
+0.213	9916_20225
+0.201	9916_20236
+0.196	9916_65281
+0.148	9916_65292
… 761 more positive …
… 49230 more negative …

Weight^?	Feature
+0.510	9900_46850
+0.445	9900_46870
+0.322	9900_46870 9900_46850
+0.313	9900_46850 9900_46870
+0.214	9900_46850 9900_46870 9900_46850
+0.209	9900_46870 9900_46850 9900_46870
+0.169	5069_51213
+0.129	5069_51210
+0.114	9900_46853
+0.107	5069_51210 5069_51213
… 691 more positive …
… 49300 more negative …

Weight^?	Feature
+0.482	7700_1427
+0.402	7700_36930
+0.313	7700_49377
+0.234	7700_49377 7700_36930
+0.214	9700_14257
+0.208	9754_15275
+0.195	7700_36930 7700_49377
+0.194	9746_16330
+0.170	7700_1427 7700_1427
+0.156	7700_36930 7700_49377 7700_36930
… 145 more positive …
… 49846 more negative …

Weight^?	Feature
+0.342	9716_18178
+0.311	9716_48389
+0.282	9716_48406
+0.266	9716_48386
+0.243	9716_7190
+0.200	9716_7170
+0.191	9732_1541
+0.181	9732_1538
+0.155	9732_19458
+0.122	9716_18178 9716_48389
… 936 more positive …
… 49055 more negative …

Weight^?	Feature
+0.473	9762_28677
+0.348	9716_13825
+0.234	9716_64516
+0.191	9716_64516 9716_13825
+0.191	5042_57847
+0.187	5077_51032
+0.178	9716_13825 9716_64516
+0.150	5047_32518
+0.135	9716_58881
+0.125	9737_36868
… 723 more positive …
… 49268 more negative …

Weight^?	Feature
+0.533	9934_11009
+0.220	9716_4097
+0.202	9716_18433
+0.199	9716_18178
+0.195	9716_4108
+0.175	9716_48389
+0.157	9701_46339
+0.142	9934_56067
+0.141	5015_63902
+0.133	5000_25376
… 769 more positive …
… 49222 more negative …

Weight^?	Feature
+0.382	9734_59906
+0.289	9716_13827
+0.283	9734_8197
+0.277	9734_35078
+0.227	9734_8194
+0.197	7716_6166
+0.175	9734_8214
+0.155	9716_64516
+0.131	9908_50689
+0.129	9934_51714
… 822 more positive …
… 49169 more negative …

Weight^?	Feature
+0.478	9935_44290
+0.419	9935_44292
+0.387	9935_44310
+0.184	9935_8449
+0.174	9935_44290 9935_44292
+0.161	9935_48385
+0.144	9935_44292 9935_44290
+0.142	9935_44292 9935_44310
+0.124	9935_44310 9935_44290
+0.119	9935_44310 9935_44292
… 634 more positive …
… 49357 more negative …

Weight^?	Feature
+0.409	7716_6434
+0.378	5010_60780
+0.336	7752_1567
+0.275	7752_1569
+0.182	5010_7588
+0.162	9752_1556
+0.154	5010_7591
+0.140	9037_56936
+0.135	7752_2087
+0.126	7752_1569 7752_1567
… 291 more positive …
… 49700 more negative …

Weight^?	Feature
+0.499	9937_59395
+0.406	9937_59424
+0.319	9937_34819
+0.258	9937_59395 9937_59424
+0.255	9937_59424 9937_59395
+0.176	9758_23298
+0.165	9937_59395 9937_59424 9937_59395
+0.159	9937_59395 9937_34819
+0.146	9937_59424 9937_59395 9937_59424
+0.138	9937_34819 9937_59395
… 543 more positive …
… 49448 more negative …

Weight^?	Feature
+0.369	9937_59394
+0.364	9937_59414
+0.319	9937_59394 9937_59414
+0.299	9937_59414 9937_59394
+0.270	9937_59414 9937_59394 9937_59414
+0.250	9937_59394 9937_59414 9937_59394
+0.190	9739_21270
+0.158	9739_21250
+0.158	9716_2052
+0.147	5024_8480
… 516 more positive …
… 49475 more negative …

Weight^?	Feature
+0.456	7787_14015
+0.354	7710_7966
+0.267	7755_6982
+0.253	9787_13978
+0.253	5006_4711
+0.249	7755_30228
+0.202	9006_4831
+0.202	9787_13978 7787_14015
+0.190	7755_6594
+0.163	7755_6593
… 69 more positive …
… 49922 more negative …

Weight^?	Feature
+0.545	7755_15802
+0.310	7755_15807
+0.303	7755_15810
+0.206	9757_7021
+0.190	7755_15806
+0.164	7755_15818
+0.162	7755_15802 7755_15807
+0.160	7755_15810 7755_15802
+0.154	7755_15802 7755_15810
+0.146	9716_33842
… 228 more positive …
… 49763 more negative …

Weight^?	Feature
+0.533	7755_6592
+0.437	7755_6976
+0.304	7755_6976 7755_6592
+0.283	7755_6592 7755_6976
+0.231	7755_6592 7755_6592
+0.204	7755_6976 7755_6976
+0.185	7755_6977
+0.155	7755_6982
+0.136	7755_6976 7755_6592 7755_6976
+0.133	7755_6592 7755_6976 7755_6976
… 119 more positive …
… 49872 more negative …

Weight^?	Feature
+0.452	9716_3665
+0.282	9716_7033
+0.267	7755_31336
+0.223	9700_19413
+0.223	9716_7033 9716_3665
+0.223	9716_3665 7755_31336
+0.175	9716_3665 9716_3665
+0.156	7755_31336 9716_3665
+0.151	9716_31386
+0.146	9716_7033 9716_3665 9716_3665
… 149 more positive …
… 49842 more negative …

Weight^?	Feature
+0.492	9916_41475
+0.463	9916_41504
+0.332	9916_41504 9916_41475
+0.332	9916_41475 9916_41504
+0.230	9916_41475 9916_41504 9916_41475
+0.215	9916_41504 9916_41475 9916_41504
+0.130	9758_23297
+0.128	9758_23308
+0.126	9916_41475 9916_41475
+0.124	9734_14854
… 727 more positive …
… 49264 more negative …

Weight^?	Feature
+0.377	7716_742
+0.373	9716_31053
+0.371	7716_3160
+0.371	9716_3151
+0.242	7716_742 9716_31053
+0.224	9716_31053 7716_742
+0.186	9716_3151 7716_3160
+0.184	7716_3160 9716_3151
+0.156	9716_31053 9716_31053
+0.132	7716_3160 7716_3160
… 272 more positive …
… 49719 more negative …

Weight^?	Feature
+0.358	9937_59394
+0.349	9937_59414
+0.349	9739_34305
+0.285	9739_9729
+0.244	9739_44294
+0.234	9937_59414 9937_59394
+0.229	9937_59394 9937_59414
+0.191	9739_9729 9739_34305
+0.169	9739_34305 9739_9729
+0.158	9937_59394 9937_59414 9937_59394
… 860 more positive …
… 49131 more negative …

Weight^?	Feature
+0.347	9942_22276
+0.342	9976_63750
+0.301	9942_22273
+0.262	9716_48918
+0.260	9942_22284
+0.256	9976_63747
+0.208	9716_48898
+0.183	9716_48918 9716_48898
+0.183	9716_48898 9716_48918
+0.155	9716_48918 9716_48898 9716_48918
… 659 more positive …
… 49332 more negative …

Weight^?	Feature
+0.548	9934_45318
+0.473	9934_257
+0.302	9934_45315
+0.216	9934_45324
+0.195	9934_45344
+0.183	9934_45318 9934_257
+0.147	9934_257 9934_45318
+0.122	9934_45324 9934_257
+0.113	9934_45315 9934_45318
+0.112	9916_25089
… 515 more positive …
… 49476 more negative …

Weight^?	Feature
+0.674	5007_64499
+0.336	9007_54879
+0.323	9007_54874
+0.214	9007_54874 5007_64499
+0.209	5007_64499 9007_54879
+0.193	5007_64499 9007_54874
+0.182	9007_54879 5007_64499
+0.174	5007_64499 5007_64499
+0.130	5007_64499 9007_54874 5007_64499
+0.122	9046_21253
… 424 more positive …
… 49567 more negative …

Weight^?	Feature
+0.407	7752_880
+0.384	7752_894
+0.303	9746_16331
+0.273	7752_880 7752_894
+0.231	9752_866
+0.228	7752_894 7752_880
+0.198	7746_48369
+0.167	7752_880 7752_894 7752_880
+0.163	9752_867
+0.152	9746_34221
… 257 more positive …
… 49734 more negative …

Weight^?	Feature
+0.473	9937_43011
+0.461	9937_43040
+0.368	9937_43011 9937_43040
+0.341	9937_43040 9937_43011
+0.257	9937_43011 9937_43040 9937_43011
+0.254	9937_43040 9937_43011 9937_43040
+0.105	9937_30722
+0.095	9937_43040 9937_43040
+0.092	9937_34817
+0.089	9937_43011 9937_43011
… 848 more positive …
… 49143 more negative …

Weight^?	Feature
+0.504	9937_30722
+0.357	9937_43040
+0.251	7716_44974
+0.246	9937_43040 9937_30722
+0.246	9937_43011
+0.216	9937_30722 9937_43040
+0.203	9937_30721
+0.152	9937_30722 9937_30722
+0.151	9937_30722 9937_43040 9937_30722
+0.141	9937_43011 9937_30722
… 752 more positive …
… 49239 more negative …

Weight^?	Feature
+0.480	7716_3251
+0.404	9716_32507
+0.330	7716_60475
+0.269	9716_32507 7716_3251
+0.233	7716_3251 9716_32507
+0.208	7752_8399
+0.188	7716_32502
+0.179	7716_60475 9716_32507
+0.129	7716_3251 7716_60475
+0.125	9716_32507 7716_60475
… 425 more positive …
… 49566 more negative …

Weight^?	Feature
+0.540	9716_56834
+0.484	9716_25356
+0.299	9716_25356 9716_56834
+0.296	9716_56834 9716_25356
+0.178	9716_56834 9716_56834
+0.152	9900_10753
+0.136	9716_25345
+0.134	9716_56834 9716_25356 9716_56834
+0.129	9716_25356 9716_56834 9716_25356
+0.127	9716_25356 9716_25356
… 500 more positive …
… 49491 more negative …

Weight^?	Feature
+0.584	7764_2759
+0.541	9764_2755
+0.248	7764_2759 9764_2755
+0.243	9764_2755 7764_2759
+0.209	7764_2770
+0.156	9764_2763
+0.137	7764_2759 9764_2755 7764_2759
+0.128	9764_2755 9764_2755
+0.126	7764_2759 7764_2759
+0.107	9764_2755 7764_2759 9764_2755
… 326 more positive …
… 49665 more negative …

Weight^?	Feature
+0.562	5022_19797
+0.534	9041_55813
+0.248	9041_36101
+0.235	5022_61065
+0.177	5022_19797 5022_19797
+0.171	9041_36098
+0.171	5022_19793
+0.153	9700_62979
+0.153	9774_50693
+0.111	7716_3672
… 19 more positive …
… 49972 more negative …

Weight^?	Feature
+0.352	9734_12832 9734_12803 9734_12832
+0.317	9734_12832 9734_12803
+0.311	9734_12803 9734_12832
+0.309	9734_12803 9734_12832 9734_12803
+0.305	9937_59393
+0.305	9937_59404
+0.259	9734_12803
+0.259	9734_12832
+0.239	9937_59404 9937_59393
+0.233	9937_59393 9937_59404
… 575 more positive …
… 49416 more negative …

Weight^?	Feature
+0.582	7716_3251
+0.342	9716_32506
+0.339	9716_32507
+0.241	7716_3251 9716_32507
+0.225	7716_3251 7716_3251
+0.220	9716_32507 7716_3251
+0.182	9716_32506 7716_3251
+0.165	7716_3251 9716_32506
+0.127	9716_32506 9716_32506
+0.123	7716_3251 7716_3251 9716_32507
… 351 more positive …
… 49640 more negative …

Weight^?	Feature
+0.483	9916_9988
+0.451	9916_15873
+0.343	9916_9986
+0.293	9916_10006
+0.209	9916_15873 9916_9988
+0.203	9916_9988 9916_15873
+0.161	9916_9986 9916_15873
+0.124	9916_9988 9916_9986
+0.121	9916_9986 9916_9988
+0.121	9916_15873 9916_9986
… 584 more positive …
… 49407 more negative …

Weight^?	Feature
+0.465	9713_40962
+0.282	9713_18692
+0.268	9739_10508
+0.264	9739_10497
+0.251	9713_6406
+0.197	9935_48386
+0.183	9739_10497 9739_10508
+0.147	9916_52228
+0.140	9739_10508 9739_10497
+0.130	9739_61189
… 637 more positive …
… 49354 more negative …

Weight^?	Feature
+0.414	9716_13058
+0.296	5000_45569
+0.261	5000_45580
+0.259	5000_45569 5000_45580
+0.237	5000_45580 5000_45569
+0.236	9716_25346
+0.222	5000_45569 5000_45580 5000_45569
+0.178	5000_45580 5000_45569 5000_45580
+0.177	9716_25346 9716_13058
+0.167	9716_13058 9716_25346
… 104 more positive …
… 49887 more negative …

Weight^?	Feature
+0.513	7700_30030
+0.332	9758_43523
+0.296	9700_30038
+0.222	9934_29445
+0.211	7700_30030 9700_30038
+0.193	9758_43521
+0.193	9934_29445 7700_30030
+0.186	7700_30030 9934_29445
+0.168	9700_30038 7700_30030
+0.159	9758_43523 9758_43521
… 660 more positive …
… 49331 more negative …

Weight^?	Feature
+0.595	5078_8747
+0.327	9052_46594
+0.244	5060_40884
+0.208	9716_37142
+0.208	5060_41625
+0.186	9716_37122
+0.179	9716_37142 9716_37122
+0.171	9716_37122 9716_37142
+0.164	9716_37142 9716_37122 9716_37142
+0.149	9716_37122 9716_37142 9716_37122
… 304 more positive …
… 49687 more negative …

Weight^?	Feature
+0.384	9701_40470
+0.373	9701_40450
+0.261	9701_40450 9701_40470
+0.244	9701_40470 9701_40450
+0.227	9059_29185
+0.220	9730_6915
+0.208	9730_34818
+0.208	9059_29187
+0.163	9739_22530
+0.162	9701_40450 9701_40470 9701_40450
… 788 more positive …
… 49203 more negative …

Weight^?	Feature
+0.367	7752_1389
+0.300	7716_31988
+0.205	9716_39954
+0.204	7753_62669
+0.183	7752_1390
+0.176	7716_36160
+0.172	7753_6136
+0.172	7730_3805
+0.162	9716_31989
+0.158	7716_60325
… 241 more positive …
… 49750 more negative …

Weight^?	Feature
+0.428	7752_1942
+0.330	7752_1960
+0.261	7752_1942 7752_1960
+0.252	7716_940
+0.245	7752_1960 7752_1942
+0.208	7716_925
+0.196	7752_1942 7752_1960 7752_1942
+0.160	7745_55705
+0.152	7752_1960 7752_1942 7752_1960
+0.145	7745_14141
… 287 more positive …
… 49704 more negative …

Weight^?	Feature
+0.445	9916_42508
+0.411	9916_42497
+0.288	9916_42508 9916_42497
+0.277	9916_42497 9916_42508
+0.269	9732_39180
+0.187	9916_42508 9916_42497 9916_42508
+0.179	9916_42497 9916_42508 9916_42497
+0.171	9732_39169
+0.165	7752_1911
+0.160	9732_39169 9732_39180
… 509 more positive …
… 49482 more negative …

Weight^?	Feature
+0.405	9734_33286
+0.342	9734_33283
+0.308	7716_3253
+0.269	9734_33312
+0.198	9734_33286 9734_33283
+0.180	9916_35841
+0.179	9734_33283 9734_33286
+0.151	9734_33312 9734_33283
+0.148	5050_63489
+0.143	9916_35852
… 791 more positive …
… 49200 more negative …

Weight^?	Feature
+0.524	9716_33832
+0.359	7716_55674
+0.310	9716_1797
+0.264	7716_55674 9716_33832
+0.256	9716_33832 7716_55674
+0.210	9716_33837
+0.174	7716_55674 9716_33832 7716_55674
+0.165	9716_33832 9716_33832
+0.145	9716_33832 9716_1797
+0.132	9716_1797 9716_33832
… 305 more positive …
… 49686 more negative …

Weight^?	Feature
+0.443	9734_47110
+0.387	9716_2265
+0.311	9716_2265 9734_47110
+0.302	9734_47110 9716_2265
+0.199	9937_30722
+0.189	9734_47110 9716_2265 9734_47110
+0.162	9937_30722 9937_30722
+0.162	9937_30722 9937_30722 9937_30722
+0.155	9937_30721
+0.145	9716_34956
… 176 more positive …
… 49815 more negative …

Weight^?	Feature
+0.438	9916_19488
+0.362	9916_19459
+0.326	9916_19459 9916_19488
+0.307	9916_19488 9916_19459
+0.231	9916_19488 9916_19459 9916_19488
+0.217	9005_26113
+0.212	9916_19459 9916_19488 9916_19459
+0.193	9754_63776
+0.180	9754_63747
+0.146	9005_46593
… 565 more positive …
… 49426 more negative …

Weight^?	Feature
+0.342	9716_44960
+0.341	7746_6325
+0.330	7745_1270
+0.266	9746_6313
+0.256	7746_6314
+0.228	9746_16332
+0.186	7745_8223
+0.186	9746_6313 7746_6314
+0.151	9736_54785
+0.150	9716_33093
… 104 more positive …
… 49887 more negative …

Weight^?	Feature
+0.573	9754_51202
+0.370	9754_51222
+0.229	9754_51202 9754_51222
+0.192	9754_51222 9754_51202
+0.182	9739_29953
+0.173	9934_6659
+0.163	9739_10500
+0.159	9739_22529
+0.159	9739_29953 9754_51202
+0.153	9754_51202 9754_51222 9754_51202
… 793 more positive …
… 49198 more negative …

Weight^?	Feature
+0.709	9716_31788
+0.237	9716_31788 9746_16332
+0.229	9746_16332 9716_31788
+0.217	9746_16332
+0.214	9716_31788 9746_16331
+0.205	9746_16331
+0.199	9746_16331 9716_31788
+0.184	9716_31788 9746_16332 9716_31788
+0.161	9716_31788 9746_16331 9716_31788
+0.158	9716_31788 9716_31788
… 170 more positive …
… 49821 more negative …

Weight^?	Feature
+0.393	7716_1843
+0.349	7716_1840
+0.306	9716_1841
+0.282	7752_1941
+0.257	7716_1842
+0.249	7752_1959
+0.191	7752_1941 7752_1959
+0.170	7752_1959 7752_1941
+0.160	5022_19540
+0.153	5022_52187
… 151 more positive …
… 49840 more negative …

Weight^?	Feature
+0.409	9716_19971
+0.344	9937_43040
+0.278	9937_43011
+0.258	9716_20000
+0.234	9716_56835
+0.218	9937_43040 9937_43011
+0.206	9937_43011 9937_43040
+0.197	9716_20000 9716_19971
+0.184	9716_19971 9716_20000
+0.154	9716_19974
… 798 more positive …
… 49193 more negative …

Weight^?	Feature
+0.323	9716_39937
+0.305	9716_39948
+0.305	9739_22531
+0.262	9716_18178
+0.198	9916_20246
+0.197	9716_39940
+0.182	9916_20226
+0.174	9716_22529
+0.140	9735_19478
+0.133	9739_22531 9716_39948
… 887 more positive …
… 49104 more negative …

Weight^?	Feature
+0.380	9739_52738
+0.345	9739_52758
+0.320	9739_52741
+0.289	9934_51714
+0.265	9934_39426
+0.199	9739_52738 9739_52758
+0.184	9739_52741 9739_52738
+0.166	9739_52758 9739_52738
+0.120	9934_43523
+0.117	9739_52738 9739_52741
… 918 more positive …
… 49073 more negative …

Weight^?	Feature
+0.314	9734_12801
+0.311	9734_12812
+0.298	9916_20226
+0.275	9916_20246
+0.234	9734_33302
+0.172	9916_20226 9916_20246
+0.172	9734_12812 9734_12801
+0.169	9916_20246 9916_20226
+0.159	9734_33282
+0.139	9916_65292
… 944 more positive …
… 49047 more negative …

Weight^?	Feature
+0.351	9916_65292
+0.317	9916_65281
+0.279	9916_20226
+0.277	9916_20246
+0.211	9716_7173
+0.190	9735_47884
+0.189	9745_10037
+0.168	9916_20246 9916_20226
+0.158	9916_20226 9916_20246
+0.139	9916_65281 9916_65292
… 1053 more positive …
… 48938 more negative …

Weight^?	Feature
+0.283	9716_7200
+0.249	9716_48406
+0.246	9916_20236
+0.223	9716_7190
+0.211	9916_20225
+0.174	9716_48386
+0.168	9716_7171
+0.163	9903_50690
+0.148	9716_7171 9716_7200
+0.145	9716_7200 9716_7171
… 773 more positive …
… 49218 more negative …

Weight^?	Feature
+0.469	9716_3144
+0.430	7716_3155
+0.283	7716_6028
+0.262	9716_3145
+0.251	7716_37505
+0.213	9716_3144 7716_3155
+0.210	7716_3155 9716_3144
+0.154	7716_3154
+0.140	9716_3145 7716_3155
+0.138	9716_3144 9716_3144
… 470 more positive …
… 49521 more negative …

Weight^?	Feature
+0.742	7716_65294
+0.271	7716_65294 7716_65294
+0.237	9716_33099
+0.196	9716_33099 7716_65294
+0.190	7716_65294 9716_33099
+0.145	7716_65294 9716_33099 7716_65294
+0.111	9716_33096
+0.109	9716_65413
+0.104	9716_33099 7716_65294 9716_33099
+0.103	7716_65294 9716_33096
… 364 more positive …
… 49627 more negative …

Weight^?	Feature
+0.371	9916_20226
+0.346	9916_20246
+0.271	9734_61702
+0.200	9916_20226 9916_20246
+0.196	9916_20246 9916_20226
+0.192	9734_61728
+0.170	9734_61699
+0.169	9916_17409
+0.136	9916_17420
+0.125	9734_16642
… 988 more positive …
… 49003 more negative …

Weight^?	Feature
+0.399	9937_43011
+0.355	9937_43040
+0.302	9937_43040 9937_43011
+0.291	9937_43011 9937_43040
+0.231	9937_43011 9937_43040 9937_43011
+0.205	9937_43040 9937_43011 9937_43040
+0.173	9909_50700
+0.152	9937_34817
+0.134	9758_11788
+0.133	9909_50689
… 962 more positive …
… 49029 more negative …

Weight^?	Feature
+0.344	9916_20225
+0.339	9916_20236
+0.288	9916_20236 9916_20225
+0.286	9716_4098
+0.274	9916_20225 9916_20236
+0.246	9716_4118
+0.225	9916_20236 9916_20225 9916_20236
+0.206	9916_20225 9916_20236 9916_20225
+0.200	9716_54038
+0.181	9043_40198
… 775 more positive …
… 49216 more negative …

Weight^?	Feature
+0.494	7752_40574
+0.337	7752_40452
+0.329	7716_55662
+0.249	7752_40454
+0.185	9716_61253
+0.168	7752_40574 7752_40454
+0.168	7752_40452 7752_40574
+0.158	7716_55662 9716_61253
+0.156	9752_40793
+0.138	7716_32952
… 261 more positive …
… 49730 more negative …

Weight^?	Feature
+0.600	9716_62011
+0.372	7716_55654
+0.273	7716_55654 9716_62011
+0.245	9716_62011 7716_55654
+0.176	9716_62011 9716_62011
+0.152	7700_33799
+0.145	9716_62011 7716_55654 9716_62011
+0.131	7716_49282
+0.128	7700_1152
+0.125	9716_62011 7716_49282
… 557 more positive …
… 49434 more negative …

Weight^?	Feature
+0.403	9701_49153
+0.268	9934_35842
+0.255	9701_49153 9934_35842
+0.254	9701_49156
+0.246	9934_35842 9701_49153
+0.200	9934_35862
+0.195	9701_49153 9934_35842 9701_49153
+0.165	9735_27660
+0.154	9701_49156 9934_35862
+0.152	9934_35842 9701_49153 9934_35842
… 780 more positive …
… 49211 more negative …

Weight^?	Feature
+0.579	7798_15404
+0.328	7798_62672
+0.218	7798_62672 7798_15404
+0.214	7755_55452
+0.197	7798_15404 7798_62672
+0.164	7798_59973
+0.159	7752_32983
+0.153	7798_34295
+0.144	7755_44696
+0.142	7798_15404 7798_15404
… 313 more positive …
… 49678 more negative …

Weight^?	Feature
+0.435	9716_30209
+0.367	9716_30220
+0.295	9716_30209 9716_30220
+0.241	9716_30220 9716_30209
+0.177	9716_30209 9716_30220 9716_30209
+0.166	9716_3841
+0.157	9716_30220 9716_30209 9716_30220
+0.120	9916_65292
+0.118	9916_65281
+0.094	9916_33025
… 1011 more positive …
… 48980 more negative …

Weight^?	Feature
+0.357	9739_29953
+0.332	9916_42753
+0.328	9916_42764
+0.268	9739_22529
+0.263	9916_42753 9916_42764
+0.240	9916_42764 9916_42753
+0.217	9754_51202
+0.209	9739_29953 9739_22529
+0.188	9916_42764 9916_42753 9916_42764
+0.186	9916_42753 9916_42764 9916_42753
… 723 more positive …
… 49268 more negative …

Weight^?	Feature
+0.343	9701_64771
+0.340	9701_64800
+0.277	9716_18433
+0.243	9701_64771 9701_64800
+0.242	9701_64800 9701_64771
+0.237	9716_4108
+0.211	9716_39942
+0.204	9716_4097
+0.201	9701_64771 9701_64800 9701_64771
+0.171	9701_64800 9701_64771 9701_64800
… 800 more positive …
… 49191 more negative …

Weight^?	Feature
+0.562	7716_902
+0.530	9734_64260
+0.279	7716_902 9734_64260
+0.270	9734_64260 7716_902
+0.218	9734_64260 9734_64260
+0.171	7716_902 7716_902
+0.137	7716_902 9734_64260 7716_902
+0.133	7716_902 9734_64260 9734_64260
+0.124	9734_64260 7716_902 9734_64260
+0.112	7716_902 7716_902 9734_64260
… 667 more positive …
… 49324 more negative …

Weight^?	Feature
+0.376	9937_38156
+0.351	9937_38145
+0.262	9937_38156 9937_38145
+0.258	9741_48901
+0.254	9937_38145 9937_38156
+0.236	7741_5928
+0.184	9014_60422
+0.180	9937_38156 9937_38145 9937_38156
+0.178	9758_34561 9758_34572
+0.177	9758_34561
… 621 more positive …
… 49370 more negative …

Weight^?	Feature
+0.476	9934_59137
+0.357	7752_59117
+0.312	9734_61702
+0.280	9734_16642
+0.277	9934_59137 7752_59117
+0.267	7752_59117 9934_59137
+0.183	9934_59137 7752_59117 9934_59137
+0.179	9734_16642 9734_61702
+0.176	9734_61702 9734_16642
+0.147	7752_59117 9934_59137 7752_59117
… 753 more positive …
… 49238 more negative …

Weight^?	Feature
+0.519	9934_56067
+0.381	9934_56076
+0.350	9934_56065
+0.197	9934_29188
+0.170	9062_51718
+0.160	9934_56076 9934_56065
+0.149	9934_56065 9934_56076
+0.146	9934_56067 9934_56076
+0.142	9934_56065 9934_56067
+0.125	9934_56067 9934_56065
… 592 more positive …
… 49399 more negative …

Weight^?	Feature
+0.440	9739_22530
+0.358	9716_2052
+0.284	9716_2052 9739_22530
+0.259	9739_22530 9716_2052
+0.251	9736_27137
+0.181	9739_22530 9716_2052 9739_22530
+0.142	9773_51468
+0.135	9739_22530 9739_22530
+0.135	9736_56070
+0.126	9773_55820
… 780 more positive …
… 49211 more negative …

Weight^?	Feature
+0.423	7755_15803
+0.409	9716_3665
+0.298	7716_3672
+0.281	7755_15811
+0.245	7755_15803 9716_3665
+0.193	9716_3665 7755_15803
+0.177	7716_3672 7755_15811
+0.172	9752_2971
+0.147	9716_33846
+0.139	7755_15811 7716_3672
… 598 more positive …
… 49393 more negative …

Weight^?	Feature
+0.439	9916_35852
+0.392	9916_35841
+0.315	9916_35841 9916_35852
+0.304	9916_35852 9916_35841
+0.299	5060_45028
+0.231	5045_45029
+0.211	5045_5896
+0.200	9916_35852 9916_35841 9916_35852
+0.200	9916_35841 9916_35852 9916_35841
+0.116	5060_45028 5045_5896
… 234 more positive …
… 49757 more negative …

Weight^?	Feature
+0.373	9916_35852
+0.309	9916_35841
+0.271	9916_35841 9916_35852
+0.265	9916_35852 9916_35841
+0.258	9754_18690
+0.214	9754_63747
+0.210	9916_35852 9916_35841 9916_35852
+0.204	9916_35841 9916_35852 9916_35841
+0.163	9754_63776
+0.149	7755_30228
… 707 more positive …
… 49284 more negative …

Weight^?	Feature
+0.562	9732_19457
+0.313	9732_39172
+0.221	9916_35841
+0.203	9732_64516
+0.200	9916_35852
+0.161	9732_1541
+0.148	9732_39172 9732_19457
+0.135	9739_34050
+0.131	9916_35841 9916_35852
+0.124	9716_43778
… 803 more positive …
… 49188 more negative …

Weight^?	Feature
+0.432	9916_35852
+0.383	9916_35841
+0.368	9916_35852 9916_35841
+0.352	9916_35841 9916_35852
+0.301	9916_35841 9916_35852 9916_35841
+0.297	9916_35852 9916_35841 9916_35852
+0.141	9716_7190
+0.125	9716_7170
+0.098	9909_49153
+0.089	9901_50945
… 926 more positive …
… 49065 more negative …

Weight^?	Feature
+0.523	9937_59393
+0.413	9937_30723
+0.349	9937_59404
+0.261	7757_7012
+0.220	9937_59393 9937_59404
+0.191	9937_30723 9937_59393
+0.177	9937_59404 9937_59393
+0.155	7757_7012 9937_30723
+0.123	9937_59393 9937_59404 9937_59393
+0.116	9937_59393 7757_7012
… 520 more positive …
… 49471 more negative …

Weight^?	Feature
+0.582	9741_24453
+0.308	7741_17689
+0.302	7741_837
+0.205	7755_872
+0.183	7752_1564
+0.171	9741_24453 7741_837
+0.168	7741_837 9741_24453
+0.168	9741_24453 9741_24453
+0.167	9741_655
+0.153	7752_1565
… 423 more positive …
… 49568 more negative …

Weight^?	Feature
+0.540	7755_7128
+0.364	7755_35929
+0.361	7755_23103
+0.338	7755_38893
+0.250	7755_38817
+0.198	7755_35948
+0.125	7755_38893 7755_7128
+0.119	7752_516
+0.111	7755_30064
+0.111	7755_7128 7755_38893
… 221 more positive …
… 49770 more negative …

Weight^?	Feature
+0.579	7716_36161
+0.371	9716_31987
+0.311	9716_31987 7716_36161
+0.298	7716_36161 7716_36161
+0.254	7716_36161 7716_36161 7716_36161
+0.246	7716_7043
+0.135	7716_7042
+0.115	7716_7043 7716_7043
+0.095	7700_36754
+0.089	7716_7043 7700_34895
… 74 more positive …
… 49917 more negative …

Weight^?	Feature
+0.486	9934_46594
+0.446	9934_46614
+0.385	9934_46614 9934_46594
+0.385	9934_46594 9934_46614
+0.331	9934_46594 9934_46614 9934_46594
+0.308	9934_46614 9934_46594 9934_46614
+0.065	9934_46594 9934_46594
+0.062	9732_40192
+0.059	9784_14614
+0.058	9934_46594 9934_46594 9934_46614
… 602 more positive …
… 49389 more negative …

Weight^?	Feature
+0.591	9752_15475
+0.316	7755_61036
+0.265	9752_15475 9752_15475
+0.228	7755_30237
+0.221	9752_15475 9752_15475 9752_15475
+0.204	7755_33346
+0.138	7755_38846
+0.128	7755_61031
+0.124	7755_33346 9700_63215 9752_14633
+0.124	9752_14633 9752_15475
… 108 more positive …
… 49883 more negative …

Weight^?	Feature
+0.623	9739_27909
+0.286	9716_1797
+0.263	9739_27909 9716_1797
+0.258	9716_1797 9739_27909
+0.231	9739_27909 9739_27909
+0.213	9716_1814
+0.184	9739_27909 9716_1814
+0.181	9716_1814 9739_27909
+0.164	9739_27909 9716_1797 9739_27909
+0.132	9716_1797 9739_27909 9716_1797
… 585 more positive …
… 49406 more negative …

Weight^?	Feature
+0.342	9934_32513
+0.241	9739_52741
+0.232	7716_3387
+0.232	9739_22530
+0.229	7716_3386
+0.215	9787_42757
+0.209	9787_42774
+0.209	9934_12037
+0.205	7755_33264
+0.188	9787_42754
… 591 more positive …
… 49400 more negative …

Weight^?	Feature
+0.645	9739_34305
+0.296	7755_6589
+0.274	9739_34305 9739_34305
+0.193	9739_34305 7755_6589
+0.193	9908_53762
+0.190	9739_61188
+0.181	9908_18947
+0.154	9739_34305 7755_6589 9739_34305
+0.145	7755_6589 9739_34305
+0.125	9739_61188 9739_61188
… 78 more positive …
… 49913 more negative …

Weight^?	Feature
+0.506	7752_3290
+0.321	7752_19475
+0.273	7752_19470
+0.246	9752_19471
+0.240	9752_1553
+0.211	7752_1564
+0.207	7745_10345
+0.199	9752_10501
+0.150	7752_55304
+0.136	7752_3290 7752_3290
… 435 more positive …
… 49556 more negative …

Weight^?	Feature
+0.455	9739_3333
+0.322	9935_53505
+0.292	9935_53516
+0.238	9935_53516 9935_53505
+0.217	9935_53505 9935_53516
+0.206	9739_3333 9716_7169
+0.187	9935_53516 9935_53505 9935_53516
+0.181	9716_7169
+0.178	9716_7169 9739_3333
+0.166	9935_53505 9935_53516 9935_53505
… 717 more positive …
… 49274 more negative …

Weight^?	Feature
+0.374	7716_3497
+0.345	9752_34072
+0.306	9716_34958
+0.301	7752_8584
+0.234	9716_34958 7716_3497
+0.201	7716_3497 9716_34958
+0.187	9752_34072 7752_8584
+0.140	7765_65407
+0.140	7758_3316
+0.140	7752_354
… 288 more positive …
… 49703 more negative …

Weight^?	Feature
+0.324	9953_37016
+0.304	9752_12949
+0.282	9716_33229
+0.236	9709_7021
+0.235	9953_37016 9716_33229
+0.189	9953_34366
+0.182	9716_34606
+0.182	9716_33229 9953_37016
+0.181	9716_34606 7730_3431
+0.164	7730_3431
… 207 more positive …
… 49784 more negative …

Weight^?	Feature
+0.499	9934_39427
+0.337	9732_63233
+0.234	9733_53254
+0.213	9934_39427 9934_45313
+0.212	9732_63233 9733_53254
+0.201	9934_45313
+0.194	9934_45313 9934_39427
+0.163	9934_39427 9934_39427
+0.145	9934_45324
+0.145	9934_45324 9934_39427
… 861 more positive …
… 49130 more negative …

Weight^?	Feature
+0.542	9754_14283
+0.477	7754_14282
+0.255	7746_16336
+0.248	9754_14283 7754_14282
+0.230	7754_14282 9754_14283
+0.177	7754_37034
+0.173	9754_14283 9754_14283
+0.168	7716_39675
+0.141	7754_14282 7754_14282
+0.131	7746_16336 7716_39675
… 319 more positive …
… 49672 more negative …

Weight^?	Feature
+0.436	9782_41740
+0.415	9782_41729
+0.332	5032_60676
+0.207	9782_41729 9782_41740
+0.199	9782_41730
+0.199	9782_41750
+0.194	9782_41740 9782_41729
+0.136	9739_29957
+0.133	9934_45313
+0.133	9782_41740 9782_41729 9782_41740
… 772 more positive …
… 49219 more negative …

Weight^?	Feature
+0.330	9734_12832
+0.286	9734_12803
+0.276	9916_62467
+0.268	9916_62496
+0.262	9916_62467 9916_62496
+0.253	9734_12832 9734_12803
+0.252	9916_62496 9916_62467
+0.249	9734_12803 9734_12832
+0.237	9734_12832 9734_12803 9734_12832
+0.222	9758_5900
… 435 more positive …
… 49556 more negative …

Weight^?	Feature
+0.474	9746_38506
+0.327	9953_6545
+0.262	9746_38506 9953_6545
+0.232	9746_38506 9746_38506
+0.221	9953_6545 9746_38506
+0.196	9709_3657
+0.190	9746_38506 9746_38506 9953_6545
+0.169	9746_38506 9953_6545 9716_33229
+0.168	9953_6545 9716_33229 7730_3431
+0.156	9953_6545 9716_33229
… 180 more positive …
… 49811 more negative …

Weight^?	Feature
+0.589	9716_33025
+0.303	9716_30211
+0.220	9716_61956
+0.216	9716_31062
+0.197	9716_16897
+0.186	7716_28965
+0.160	7716_28965 9716_33025
+0.143	9716_33025 7716_28965
+0.137	9716_33025 9716_33025
+0.125	9935_14338
… 667 more positive …
… 49324 more negative …

Weight^?	Feature
+0.504	9716_1841
+0.449	7716_1843
+0.311	7716_1840
+0.298	7716_1843 9716_1841
+0.279	9716_1841 7716_1843
+0.189	9716_1841 7716_1840
+0.175	7716_1843 9716_1841 7716_1843
+0.134	7716_1840 9716_1841
+0.129	9716_1841 7716_1843 9716_1841
+0.103	7716_1843 7716_1843
… 180 more positive …
… 49811 more negative …

Weight^?	Feature
+0.405	9700_730
+0.390	9754_3559
+0.333	9716_33837
+0.322	9754_44237
+0.208	9745_10308
+0.160	9745_47196
+0.146	9700_10537
+0.142	9754_3559 9700_730
+0.139	9754_44237 9700_10537
+0.137	9716_7033
… 211 more positive …
… 49780 more negative …

Weight^?	Feature
+0.369	7755_41965
+0.272	7755_41961
+0.263	9752_14633 9752_19184
+0.263	9752_19184
+0.213	9953_34466
+0.211	9752_19184 9700_8607
+0.211	9700_32785 9752_14633 9752_19184
+0.211	9752_14633 9752_19184 9700_8607
+0.197	7762_32289
+0.184	9752_19184 9700_8607 7755_32379
… 89 more positive …
… 49902 more negative …

Weight^?	Feature
+0.426	9716_61259
+0.400	9716_7033
+0.310	9716_7033 9716_61259
+0.271	9716_7033 9716_7033 9716_61259
+0.237	9716_7033 9716_7033
+0.232	9752_34077
+0.232	9716_61259 9700_32789
+0.159	9700_8607 9716_7033
+0.155	7717_39608
+0.133	9700_8607
… 124 more positive …
… 49867 more negative …

Weight^?	Feature
+0.657	9716_36734
+0.318	9716_36732 9716_36734
+0.298	9716_36734 9716_36732
+0.297	9716_36732
+0.258	9716_36734 9716_36732 9716_36734
+0.231	9716_36734 9716_36734
+0.179	9716_36732 9716_36734 9716_36732
+0.119	9716_7033 9716_36734
+0.116	9752_13277
+0.112	9752_13510 9752_13277
… 79 more positive …
… 49912 more negative …

Weight^?	Feature
+0.420	7755_31336
+0.271	7755_31336 7755_31336
+0.241	9700_14123 9716_7033
+0.225	9700_14123 9716_7033 7755_31336
+0.198	9797_12439
+0.197	9716_7033 7755_31336
+0.197	9754_10667 9700_14123 9716_7033
+0.184	9700_14123
+0.183	9716_6428
+0.183	9797_12439 9754_10667 9700_14123
… 65 more positive …
… 49926 more negative …

Weight^?	Feature
+0.296	9746_31436
+0.248	9716_12396
+0.226	9752_3089
+0.222	9716_44955 9746_31436
+0.221	9716_44955
+0.216	9700_29750 9752_30046
+0.201	9752_30046 9752_3089
+0.201	9716_34958 9700_29750 9752_30046
+0.201	9716_34958 9700_29750
+0.198	9746_31436 9716_34958
… 123 more positive …
… 49868 more negative …

Weight^?	Feature
+0.382	7755_32046
+0.304	9745_10306
+0.284	7755_32046 7755_32046
+0.282	9047_46629
+0.282	7761_30007
+0.265	9745_10306 9745_10306
+0.265	9745_36247
+0.212	9745_10306 9745_36247
+0.212	9745_10306 9745_10306 9745_10306
+0.131	9745_12668
… 52 more positive …
… 49939 more negative …

Weight^?	Feature
+0.394	9716_33836
+0.380	9752_10379
+0.221	9752_10371 9752_10379
+0.201	9716_33838
+0.199	9716_7033 9716_33836
+0.177	9700_8607 9716_7033 9716_33836
+0.177	9752_10379 9752_10371 9752_10379
+0.166	9752_10379 9752_10371
+0.155	9716_7033 9716_33836 9716_33836
+0.155	9752_10379 9742_25596
… 170 more positive …
… 49821 more negative …

Weight^?	Feature
+0.517	9716_32507
+0.248	9716_32507 9716_32507
+0.197	9700_33819 9752_14633
+0.172	9716_35306 9716_32507 9716_32507
+0.172	9716_35306 9716_32507
+0.168	9752_46818
+0.162	9700_33819 9752_14633 9752_46818
+0.158	9752_14633 9752_46818
+0.154	5018_53966
+0.139	9700_33819
… 195 more positive …
… 49796 more negative …

Weight^?	Feature
+0.369	9716_37508
+0.317	5069_61001
+0.286	5069_60963
+0.268	5022_58668
+0.239	5069_53248
+0.205	5022_55587
+0.205	5022_52826
+0.183	7755_38926
+0.179	5018_59584 5022_52826
+0.158	5018_57337
… 56 more positive …
… 49935 more negative …

Weight^?	Feature
+0.595	9716_31786
+0.265	9005_45062
+0.239	9716_44950
+0.233	9716_44950 9716_31786
+0.219	9716_31786 9716_31786
+0.206	9700_15158
+0.204	9716_31786 9716_44950
+0.187	9716_31786 9716_44950 9716_31786
+0.171	9005_40997
+0.171	9005_44359
… 157 more positive …
… 49834 more negative …

Weight^?	Feature
+0.523	9716_33096
+0.276	9716_31386 9716_31386
+0.268	9716_31386 9716_31386 9716_31386
+0.219	9716_31386
+0.216	9716_900
+0.190	9716_34958
+0.151	9716_900 9716_33096
+0.144	9752_46818 9700_1655
+0.144	9716_33096 9747_34926
+0.144	9716_33096 9747_34926 9702_15768
… 83 more positive …
… 49908 more negative …

Weight^?	Feature
+0.528	9716_35038
+0.306	9752_1375
+0.254	9716_35306
+0.196	9716_35038 9700_33819
+0.196	9716_35306 9716_35038
+0.196	9752_14477 9752_1375
+0.196	9716_31386 9716_35038
+0.171	9716_31386
+0.168	9700_33818
+0.164	9716_35038 9700_33818
… 65 more positive …
… 49926 more negative …

Weight^?	Feature
+0.394	9752_31489 9702_30578
+0.391	9702_30578
+0.365	9702_30578 9752_31489
+0.359	9702_30578 9752_31489 9702_30578
+0.307	9752_31489
+0.294	9752_31489 9702_30578 9752_31489
+0.168	9746_16333
+0.139	9716_33398
+0.131	9747_15282 9716_33398
+0.120	9747_15282
… 61 more positive …
… 49930 more negative …

Weight^?	Feature
+0.794	9752_36583
+0.340	9752_36586
+0.263	9716_32508
+0.172	9716_33092
+0.148	9716_44963
+0.130	9752_43900
+0.097	9700_63226
+0.092	9752_15548 9752_11327
+0.085	9716_44963 9716_44963
+0.085	9702_36257
… 17 more positive …
… 49974 more negative …

Weight^?	Feature
+0.471	7755_37278
+0.403	9716_3664 7755_37278
+0.391	9716_3664
+0.386	7755_37278 9716_3664
+0.273	7755_37278 9716_3664 7755_37278
+0.240	9716_3664 7755_37278 9716_3664
+0.173	7755_37278 7755_37278
+0.159	7755_37278 7755_37278 9716_3664
+0.157	9716_3664 7755_37278 7755_37278
+0.139	9716_3664 9716_3664
… 74 more positive …
… 49917 more negative …

Weight^?	Feature
+0.625	9716_33376
+0.287	9716_33376 9716_33376
+0.239	9752_30046
+0.223	9716_31386
+0.140	9700_33818
+0.132	9013_9170
+0.131	9028_65017
+0.131	9700_33818 9752_30046
+0.108	9028_41996 9013_9170
+0.108	9752_30046 9700_1655
… 99 more positive …
… 49892 more negative …

Weight^?	Feature
+0.377	9746_34226
+0.346	9746_14891
+0.318	9746_14876
+0.256	9746_34226 9716_34958
+0.240	9752_28036
+0.220	9746_14876 9746_34226
+0.219	9746_14891 9746_14891
+0.216	9716_34958
+0.183	9746_14891 9746_34226
+0.146	9746_14891 9746_14891 9746_34226
… 105 more positive …
… 49886 more negative …

Weight^?	Feature
+0.336	9716_3664
+0.265	9752_836
+0.247	9752_46818
+0.241	9716_3665
+0.207	9752_46818 9752_836
+0.204	7755_31336
+0.196	7755_31336 9716_3664
+0.190	9752_836 9752_46818
+0.161	9752_46818 9752_836 9752_46818
+0.161	9752_836 9752_46818 9752_836
… 238 more positive …
… 49753 more negative …

Weight^?	Feature
+0.493	9716_33846
+0.351	9716_3665
+0.280	9716_3665 9716_33846
+0.279	9716_61257
+0.245	9716_33846 9716_3665
+0.197	9752_2902
+0.191	9752_2902 9752_2904
+0.191	9752_2904 9752_2902
+0.179	9752_2904 9752_2902 9752_2904
+0.165	9752_2904
… 86 more positive …
… 49905 more negative …

Weight^?	Feature
+0.485	7753_62669
+0.442	7753_6136
+0.311	7716_26290
+0.260	7716_28801
+0.247	7716_26292
+0.160	7716_36160
+0.148	7716_31984
+0.146	9716_26295
+0.107	7753_62669 7753_62669
+0.102	7753_62669 7716_26292
… 570 more positive …
… 49421 more negative …

Weight^?	Feature
+0.488	9716_33099
+0.404	9716_65414
+0.403	9752_19186
+0.322	5086_54588
+0.319	9716_34958
+0.215	9716_33099 9752_19186
+0.192	9716_65414 9716_33099
+0.191	9716_34958 9716_34958
+0.178	9716_65413
+0.130	9716_33093
… 14 more positive …
… 49977 more negative …

Weight^?	Feature
+0.525	9701_2817
+0.446	9701_40450
+0.354	9701_40470
+0.263	9701_40450 9701_2817
+0.251	9701_2817 9701_40450
+0.186	9701_2817 9701_40470
+0.170	9701_40470 9701_2817
+0.160	9701_2817 9701_2817
+0.124	9701_2817 9701_40450 9701_2817
+0.104	9701_40450 9701_40450
… 428 more positive …
… 49563 more negative …

Weight^?	Feature
+0.618	9757_9091
+0.488	9953_6545
+0.269	7743_9090
+0.228	9757_9090
+0.205	9700_15158
+0.170	9953_6544
+0.159	9953_6545 9953_6545
+0.146	9953_6544 9953_6544
+0.136	9700_34898
+0.123	9716_44950
… 16 more positive …
… 49975 more negative …

Weight^?	Feature
+0.394	9746_38746
+0.323	9746_38746 9746_38746
+0.263	9716_44955
+0.227	9746_14892 9716_44955
+0.222	9746_14892
+0.195	9716_2260 9746_38746
+0.189	9746_14891
+0.180	9716_44955 9700_33819
+0.161	9746_38746 9746_38746 9746_38746
+0.161	9746_14892 9716_44955 9700_33819
… 114 more positive …
… 49877 more negative …

Weight^?	Feature
+0.409	7752_1567
+0.377	9752_1556
+0.298	9935_48641
+0.247	9935_22533
+0.223	7752_1567 7752_1567
+0.211	9752_1556 7752_1567
+0.199	7752_1567 9752_1556
+0.195	9752_1556 9752_1556
+0.158	9716_25376
+0.153	9941_9220
… 425 more positive …
… 49566 more negative …

Weight^?	Feature
+0.404	9752_13277
+0.384	9752_2446
+0.364	9752_13277 9752_2446
+0.298	9752_13277 9752_13277
+0.298	9752_2446 9752_13277
+0.273	9752_13277 9752_13277 9752_2446
+0.273	9752_13277 9752_2446 9752_13277
+0.227	9752_2446 9752_13277 9752_13277
+0.221	9716_37501
+0.209	9716_36733
… 26 more positive …
… 49965 more negative …

Weight^?	Feature
+0.422	9734_30476
+0.378	9734_30465
+0.285	9734_30465 9734_30476
+0.266	9734_30476 9734_30465
+0.240	9734_30476 9734_30465 9734_30476
+0.236	9734_30465 9734_30476 9734_30465
+0.229	9734_30467
+0.181	9916_20246
+0.180	9916_20226
+0.126	9916_20226 9916_20246
… 627 more positive …
… 49364 more negative …

Weight^?	Feature
+0.527	9701_34049
+0.340	9701_64774
+0.239	9701_46337
+0.231	9734_1538
+0.207	9701_34049 9701_64774
+0.205	9701_21506
+0.162	9701_64771
+0.148	9701_64800
+0.145	9734_2821
+0.127	9701_64774 9701_34049
… 331 more positive …
… 49660 more negative …

Weight^?	Feature
+0.492	9754_18690
+0.367	9937_40193
+0.282	9937_28162
+0.212	9937_28161
+0.187	9754_63747
+0.186	9754_63776
+0.170	9754_63750
+0.161	9754_51202
+0.140	9754_51202 9754_18690
+0.116	9754_51222
… 741 more positive …
… 49250 more negative …

Weight^?	Feature
+0.539	9716_31428
+0.358	9716_33096
+0.190	9746_38746
+0.167	9716_31428 9716_31428
+0.125	9746_38746 9747_39227
+0.119	9716_33097
+0.118	9716_31428 9716_2264
+0.118	9764_20059 9743_23276 9757_930
+0.112	9743_23276 9757_930
+0.109	9716_2264
… 147 more positive …
… 49844 more negative …

Weight^?	Feature
+0.358	7755_32046
+0.333	9752_19188
+0.318	7755_32379
+0.298	9752_14633 9752_19188
+0.188	9700_8607 7755_32379
+0.186	9700_30157 9752_14633 9752_19188
+0.170	9700_30157 9752_14633
+0.169	9752_14633
+0.149	7755_32046 7755_32046
+0.149	9752_19188 9700_8607
… 123 more positive …
… 49868 more negative …

Weight^?	Feature
+0.362	9716_33026
+0.354	9716_1797
+0.267	9716_30212
+0.180	9716_20649
+0.146	9734_30468
+0.144	9716_1797 9716_30212
+0.139	9716_1798
+0.136	9716_56834
+0.131	9763_12037
+0.128	9734_5636
… 793 more positive …
… 49198 more negative …

Weight^?	Feature
+0.429	7755_46910
+0.372	7755_46890
+0.312	7755_63261
+0.208	7757_3563
+0.208	9733_14809
+0.202	9757_3551
+0.197	7752_515
+0.163	7752_516
+0.156	7755_63261 7755_46910
+0.125	7733_14813
… 277 more positive …
… 49714 more negative …

Weight^?	Feature
+0.292	9734_59905
+0.269	9916_42753
+0.240	9916_42764
+0.239	9734_5644
+0.189	9734_30466
+0.185	9734_5644 9734_5644
+0.163	7746_14897
+0.162	9734_59905 9734_59905
+0.156	5061_1721
+0.154	9734_47105
… 291 more positive …
… 49700 more negative …

Weight^?	Feature
+0.438	7725_30574
+0.309	9716_2265
+0.308	7716_2278
+0.300	7716_2269
+0.215	7716_2284
+0.208	9700_30038
+0.192	7716_2284 7716_2278
+0.187	9716_34956
+0.186	7716_2278 7716_2284
+0.158	7725_30574 9700_30038
… 361 more positive …
… 49630 more negative …

Weight^?	Feature
+0.447	7755_31336
+0.314	9716_3664
+0.285	7755_31336 9716_3664
+0.250	9716_3666
+0.204	7755_2984
+0.192	7755_2985 9716_3666
+0.175	7755_31336 7755_31336
+0.160	7755_2985 9716_3666 7755_31336
+0.160	9752_14641 9700_8607 9716_7033
+0.150	9752_31488 9752_14641
… 116 more positive …
… 49875 more negative …

Weight^?	Feature
+0.320	7755_32767
+0.261	9716_7033 7755_33266
+0.261	9716_6428 9700_32789
+0.261	9716_6428 9700_32789 9752_14633
+0.224	7755_32269
+0.221	9716_7033
+0.210	7755_32767 9716_6428
+0.198	9716_6428
+0.186	7755_32768
+0.179	7755_33266
… 97 more positive …
… 49894 more negative …

Weight^?	Feature
+0.623	7701_5213
+0.487	9701_40453
+0.278	9701_40453 7701_5213
+0.244	7701_5213 9701_40453
+0.170	7701_5213 7701_5213
+0.164	7710_38912
+0.142	7701_5213 9701_40453 7701_5213
+0.140	7716_49282
+0.136	9701_40453 7701_5213 9701_40453
+0.112	9710_4101
… 268 more positive …
… 49723 more negative …

Weight^?	Feature
+0.402	9710_32513
+0.360	9710_13569
+0.339	9710_32524
+0.239	9710_58636
+0.230	9710_58625
+0.209	5064_38401
+0.180	5064_38412
+0.173	9734_14852
+0.155	9723_33538
+0.133	7740_2590
… 463 more positive …
… 49528 more negative …

Weight^?	Feature
+0.600	7752_6312
+0.323	7752_6312 7752_6312
+0.211	9758_43523
+0.211	7700_30030
+0.179	7752_6312 9752_14642
+0.162	9752_14642 7752_6312
+0.153	7725_30574
+0.150	7752_6312 7752_6312 7752_6312
+0.143	9752_14642
+0.137	7700_8608
… 613 more positive …
… 49378 more negative …

Weight^?	Feature
+0.358	9716_33026
+0.342	9716_43779
+0.339	9916_20226
+0.294	9916_20246
+0.203	9916_20226 9916_20246
+0.199	9916_20246 9916_20226
+0.158	9916_20225
+0.156	9916_20236
+0.155	9916_20226 9916_20246 9916_20226
+0.150	9946_25345
… 787 more positive …
… 49204 more negative …

Weight^?	Feature
+0.482	9916_41484
+0.479	9916_41473
+0.357	9916_41484 9916_41473
+0.333	9916_41473 9916_41484
+0.256	9916_41484 9916_41473 9916_41484
+0.252	9916_41473 9916_41484 9916_41473
+0.132	9739_34070
+0.125	9739_34050
+0.115	9739_34050 9739_34070 9739_34050
+0.106	9739_34070 9739_34050
… 706 more positive …
… 49285 more negative …

Weight^?	Feature
+0.433	9734_47136
+0.399	9734_47107 9734_47136
+0.397	9734_47136 9734_47107
+0.394	9734_47107
+0.294	9734_47136 9734_47107 9734_47136
+0.290	9734_47107 9734_47136 9734_47107
+0.148	9734_47136 9734_47107 9734_47107
+0.140	9734_47136 9734_47136
+0.139	9734_47107 9734_47136 9734_47136
+0.138	9734_47107 9734_47107
… 218 more positive …
… 49773 more negative …

Weight^?	Feature
+0.528	7716_2284
+0.417	9716_34957
+0.286	9716_2265
+0.249	7716_2278
+0.245	9716_34957 7716_2284
+0.168	7716_2284 9716_34957
+0.147	7716_2269
+0.145	7716_2283
+0.120	7716_2284 7716_2278
+0.103	9716_2265 7716_2284
… 270 more positive …
… 49721 more negative …

Weight^?	Feature
+0.558	7746_16339
+0.312	7746_16345
+0.230	7716_55676
+0.168	7716_31789
+0.155	7742_46754
+0.135	7716_31789 7746_16345
+0.124	7716_55676 7746_16339
+0.119	7746_16353
+0.116	7746_16339 7716_55676
+0.116	9701_43536
… 434 more positive …
… 49557 more negative …

Weight^?	Feature
+0.504	9916_51458
+0.343	9916_35841
+0.319	9790_55809
+0.242	9734_35329
+0.191	9735_13060
+0.155	9916_15366
+0.151	9794_28417
+0.140	9916_51458 9916_35841
+0.132	9935_53250
+0.132	9916_32770
… 626 more positive …
… 49365 more negative …

Weight^?	Feature
+0.483	9935_59936
+0.463	9935_59907
+0.342	9935_59936 9935_59907
+0.300	9935_59907 9935_59936
+0.225	9935_59907 9935_59936 9935_59907
+0.210	9935_59936 9935_59907 9935_59936
+0.191	9934_40963
+0.128	9734_14860
+0.117	9935_59936 9935_59936
+0.110	9935_59907 9935_59907
… 661 more positive …
… 49330 more negative …

Weight^?	Feature
+0.555	9916_65026
+0.467	9916_65046
+0.341	9916_65028
+0.248	9916_65026 9916_65046
+0.222	9916_65046 9916_65026
+0.155	9916_65046 9916_65028
+0.149	9916_65028 9916_65026
+0.124	9916_65028 9916_65046
+0.120	9916_65026 9916_65028
+0.113	9916_65026 9916_65026
… 689 more positive …
… 49302 more negative …

Weight^?	Feature
+0.469	9916_19457
+0.456	9916_19468
+0.267	9716_1794
+0.261	9916_19457 9916_19468
+0.241	9916_19468 9916_19457
+0.234	9716_1814
+0.187	7752_32993
+0.173	9716_1794 9716_1814
+0.149	9916_19457 9916_19468 9916_19457
+0.146	9716_1814 9716_1794
… 727 more positive …
… 49264 more negative …

Weight^?	Feature
+0.402	9702_30576
+0.390	9702_30576 9752_14642
+0.326	9752_14642 9702_30576
+0.315	9752_14642
+0.276	9752_14642 9702_30576 9752_14642
+0.258	9702_30576 9752_14642 9702_30576
+0.192	9716_32507
+0.168	9716_20259
+0.162	9702_30571
+0.152	9716_20259 9716_32507
… 224 more positive …
… 49767 more negative …

Weight^?	Feature
+0.302	9770_10158
+0.283	7798_19230
+0.204	9716_31063
+0.203	7762_30257
+0.172	9767_14249
+0.166	9767_14244
+0.157	7765_30668
+0.156	7762_32129
+0.155	9702_30571
+0.151	9752_11327
… 220 more positive …
… 49771 more negative …

Weight^?	Feature
+0.447	9716_18433
+0.409	9716_39968
+0.380	9716_39942
+0.376	9716_39939
+0.177	9716_18433 9716_39942
+0.155	9716_39942 9716_18433
+0.134	9716_39968 9716_18433
+0.132	9716_18433 9716_39939
+0.122	9716_48386
+0.118	9716_39939 9716_18433
… 709 more positive …
… 49282 more negative …

Weight^?	Feature
+0.394	9934_39426
+0.266	9739_24326
+0.240	9976_3872
+0.206	9934_51714
+0.199	9976_24322
+0.195	9976_3843
+0.183	9716_7180
+0.166	9976_3846
+0.163	9716_7169
+0.158	9909_16133
… 905 more positive …
… 49086 more negative …

Weight^?	Feature
+0.558	9953_37016
+0.380	9716_33102
+0.311	9953_37016 9716_33102
+0.304	9716_33102 9953_37016
+0.213	7716_26290
+0.199	9716_26295
+0.178	9953_37016 9716_33102 9953_37016
+0.147	9716_33102 9953_37016 9716_33102
+0.136	9953_37016 9716_26295
+0.129	9953_37016 9953_37016
… 279 more positive …
… 49712 more negative …

Weight^?	Feature
+0.487	7752_4063
+0.409	9702_11009
+0.238	9702_11009 7752_4063
+0.238	7752_4063 9702_11009
+0.191	7702_36002
+0.159	7752_4063 7752_4063
+0.158	7702_42622
+0.157	7755_6590
+0.148	7755_6591
+0.143	9702_11009 9702_11009
… 223 more positive …
… 49768 more negative …

Weight^?	Feature
+0.417	7700_5872
+0.347	7752_6312
+0.345	7700_30030
+0.276	9934_29445
+0.209	9700_30038
+0.166	9934_29445 7700_30030
+0.149	7752_6312 7700_30030
+0.138	7700_59709
+0.136	7700_30030 7752_6312
+0.135	7752_6312 7752_6312
… 284 more positive …
… 49707 more negative …

Weight^?	Feature
+0.539	7716_3323
+0.345	9716_33102
+0.334	9716_33102 7716_3323
+0.322	7716_3323 9716_33102
+0.256	7716_3323 7716_3323
+0.209	7716_3323 9716_33102 7716_3323
+0.188	9716_33229 7716_3323
+0.170	9716_33229
+0.166	9716_33102 7716_3323 9716_33102
+0.144	7716_3323 9716_33229
… 275 more positive …
… 49716 more negative …

Weight^?	Feature
+0.423	9701_47361
+0.399	9701_47372
+0.280	9739_10501
+0.271	9701_47361 9701_47372
+0.267	9701_47372 9701_47361
+0.249	9701_25601
+0.189	9701_47361 9701_47372 9701_47361
+0.175	9739_56580
+0.156	9701_47372 9701_47361 9701_47372
+0.145	9701_25612
… 479 more positive …
… 49512 more negative …

Weight^?	Feature
+0.333	7755_37276
+0.283	7755_3729
+0.265	7754_37010
+0.246	7755_306
+0.234	7752_3243
+0.223	7755_55452
+0.210	7754_5721
+0.177	7754_37010 7754_5721
+0.177	7754_5721 7754_37010
+0.140	9752_31486
… 140 more positive …
… 49851 more negative …

Weight^?	Feature
+0.438	9934_56067
+0.361	9934_56065
+0.349	9934_56076
+0.310	9934_11009
+0.216	9716_15874
+0.203	9716_15894
+0.195	9716_54048
+0.155	9934_56076 9934_56065
+0.138	9934_56065 9934_56076
+0.135	9934_56067 9934_11009
… 697 more positive …
… 49294 more negative …

Weight^?	Feature
+0.451	9716_48396
+0.373	9716_18177
+0.271	9716_48388
+0.265	9716_48385
+0.216	9716_48396 9716_48385
+0.190	9716_48385 9716_48396
+0.177	9716_18177 9716_48396
+0.150	9716_48388 9716_18177
+0.134	9716_48396 9716_48388
+0.132	9716_48396 9716_48396
… 783 more positive …
… 49208 more negative …

Weight^?	Feature
+0.323	9734_39948
+0.292	9734_39937
+0.254	9937_33804
+0.250	9934_56076
+0.235	9937_33793
+0.226	9734_64262
+0.211	9934_11009
+0.184	9934_56067
+0.169	9734_48387
+0.167	9782_22550
… 611 more positive …
… 49380 more negative …

Weight^?	Feature
+0.420	9735_13068
+0.295	9735_13057
+0.269	9735_13060
+0.228	7746_6552
+0.215	9758_55564
+0.209	9953_6544
+0.189	9937_33794
+0.182	9937_33814
+0.154	9937_1538
+0.150	9735_13068 9735_13057
… 863 more positive …
… 49128 more negative …

Weight^?	Feature
+0.451	9734_10006
+0.435	9734_9986
+0.361	9734_10006 9734_9986
+0.355	9734_9986 9734_10006
+0.277	9734_10006 9734_9986 9734_10006
+0.269	9734_9986 9734_10006 9734_9986
+0.140	9734_9985
+0.129	9734_10006 9734_10006
+0.114	9734_9986 9734_9986
+0.103	9734_9986 9734_9986 9734_10006
… 340 more positive …
… 49651 more negative …

Weight^?	Feature
+0.372	7700_30030
+0.274	7700_5871
+0.253	7755_6988
+0.229	7752_6312
+0.223	7700_5872
+0.192	7755_55462
+0.176	7700_59709
+0.162	7752_6312 7700_30030
+0.160	7700_30030 7752_6312
+0.150	5091_64480
… 454 more positive …
… 49537 more negative …

Weight^?	Feature
+0.494	9790_55809
+0.345	9916_51458
+0.338	7752_8398
+0.290	9785_52993
+0.228	9794_28417
+0.177	9935_53250
+0.176	9716_39425
+0.168	9716_39425 9785_52993
+0.153	9916_32770
+0.145	9785_52993 9716_39425
… 504 more positive …
… 49487 more negative …

Weight^?	Feature
+0.470	9935_55052
+0.392	9935_55041
+0.293	9935_55041 9935_55052
+0.281	9935_55052 9935_55041
+0.251	9935_64260
+0.233	9935_55041 9935_55052 9935_55041
+0.228	9935_55052 9935_55041 9935_55052
+0.181	9935_55044
+0.138	9754_63747
+0.129	7716_3669
… 552 more positive …
… 49439 more negative …

Weight^?	Feature
+0.461	9937_63500
+0.414	9937_63489
+0.338	9937_63500 9937_63489
+0.305	9937_63489 9937_63500
+0.228	9937_63500 9937_63489 9937_63500
+0.220	9937_63489 9937_63500 9937_63489
+0.193	9739_27906
+0.165	9739_27926
+0.139	9739_27906 9739_27926
+0.139	9739_27926 9739_27906
… 692 more positive …
… 49299 more negative …

Weight^?	Feature
+0.467	9701_40460
+0.321	9716_55556
+0.285	9716_55558
+0.285	9701_40449
+0.217	9716_10497
+0.170	9700_813
+0.166	9701_40449 9701_40460
+0.161	9701_40460 9701_40449
+0.153	9716_55556 9716_55558
+0.141	9701_40451
… 771 more positive …
… 49220 more negative …

Weight^?	Feature
+0.641	9746_16333
+0.277	9716_31787
+0.243	9746_16333 9746_16333
+0.229	9716_31787 9746_16333
+0.197	9746_16333 9716_31787
+0.171	9752_15568
+0.169	9752_15916
+0.162	9716_31788
+0.139	9752_15568 9752_15916
+0.139	9716_31788 9746_16333
… 100 more positive …
… 49891 more negative …

Weight^?	Feature
+0.333	9734_47107
+0.306	9734_47136
+0.277	9736_43788
+0.275	9736_43777
+0.255	9734_47136 9734_47107
+0.231	9734_47107 9734_47136
+0.210	9734_47110
+0.200	9734_47107 9734_47136 9734_47107
+0.178	9734_47136 9734_47107 9734_47136
+0.178	9736_43788 9736_43777
… 618 more positive …
… 49373 more negative …

Weight^?	Feature
+0.451	7755_6589
+0.286	7755_6590
+0.273	7755_6590 7755_6589
+0.254	7757_46774
+0.250	7755_6589 7755_6590
+0.244	7755_6589 7755_6590 7755_6589
+0.198	7767_10391
+0.184	7755_6591
+0.157	7755_6589 7755_6591
+0.142	7755_6590 7755_6589 7755_6590
… 472 more positive …
… 49519 more negative …

Weight^?	Feature
+0.454	9004_56331
+0.341	9004_54141
+0.249	7752_7749
+0.234	5046_54144
+0.204	7716_35363
+0.195	9004_54141 9004_56331
+0.182	9004_56331 9004_54141
+0.171	9716_20251
+0.168	7716_35365
+0.160	9752_11191
… 352 more positive …
… 49639 more negative …

Weight^?	Feature
+0.563	9934_22274
+0.320	9934_29185
+0.302	9739_10500
+0.215	9934_22274 9934_29185
+0.210	9934_30467
+0.194	7752_8214
+0.191	9934_29185 9934_22274
+0.168	9934_22274 9934_22274
+0.128	9934_22274 7752_8214
+0.128	7752_8214 9934_22274
… 682 more positive …
… 49309 more negative …

Weight^?	Feature
+0.500	9916_15366
+0.345	7752_10367
+0.341	7752_8398
+0.212	9916_3333
+0.188	9752_11327
+0.187	7752_10367 9916_15366
+0.176	9916_15366 7752_10367
+0.162	7752_15672
+0.125	9916_15366 9916_15366
+0.114	9916_3333 9916_15366
… 650 more positive …
… 49341 more negative …

Weight^?	Feature
+0.429	9935_64278
+0.413	9935_64258
+0.322	9935_64278 9935_64258
+0.302	9935_64258 9935_64278
+0.224	9935_55062
+0.216	9935_64258 9935_64278 9935_64258
+0.180	9935_55042
+0.177	9935_64278 9935_64258 9935_64278
+0.175	9935_64278 9935_55062
+0.153	9935_55062 9935_64278
… 793 more positive …
… 49198 more negative …

Weight^?	Feature
+0.407	7752_7732
+0.400	7745_3088
+0.297	7745_3088 7752_7732
+0.294	7752_7732 7745_3088
+0.218	7752_10505
+0.200	7752_7732 7752_7732
+0.192	7745_3089
+0.191	7745_3088 7745_3088
+0.179	7752_10505 7745_3089
+0.176	7752_7732 7745_3088 7752_7732
… 252 more positive …
… 49739 more negative …

Weight^?	Feature
+0.554	7745_46879
+0.390	7745_46876
+0.272	7752_181
+0.269	7752_182
+0.256	7752_183
+0.216	7752_180
+0.140	7745_46879 7752_183
+0.135	7745_40256
+0.131	7752_182 7745_46876
+0.119	7752_183 7745_46879
… 424 more positive …
… 49567 more negative …

Weight^?	Feature
+0.423	9916_1793
+0.332	9916_1804
+0.262	9935_44803
+0.256	9716_39425
+0.236	7752_13090
+0.234	9935_44801
+0.216	9916_1804 9916_1793
+0.209	9916_1793 9916_1804
+0.184	7752_9957
+0.184	9935_44812
… 594 more positive …
… 49397 more negative …

Weight^?	Feature
+0.609	9953_6547
+0.307	9953_6547 9716_33229
+0.288	9716_33229
+0.247	9953_6547 9953_6547
+0.227	9953_6547 9953_6547 9716_33229
+0.151	9953_45296
+0.146	9953_39591
+0.121	9953_32906
+0.110	9702_30578
+0.109	9752_31489
… 119 more positive …
… 49872 more negative …

Weight^?	Feature
+0.570	9790_29186
+0.450	9790_5890
+0.293	9790_5890 9790_29186
+0.291	9790_29186 9790_5890
+0.195	9790_5910
+0.183	9790_29186 9790_29186
+0.127	9790_29186 9790_5890 9790_29186
+0.121	9790_5890 9790_5890
+0.117	7742_473
+0.115	9790_5890 9790_29186 9790_5890
… 467 more positive …
… 49524 more negative …

Weight^?	Feature
+0.400	9025_58026
+0.357	7700_55154
+0.292	9074_51603
+0.231	5056_18948
+0.217	7700_5871
+0.212	9934_29445
+0.169	5056_16385
+0.161	9705_46338
+0.136	9701_49154
+0.131	5074_64850
… 787 more positive …
… 49204 more negative …

Weight^?	Feature
+0.439	9716_33846
+0.335	9716_33841
+0.285	9757_7016
+0.277	5069_59237
+0.248	9752_1936
+0.189	9757_9090
+0.185	5069_60912
+0.173	9716_33846 9716_33841
+0.159	9716_33846 9716_33846
+0.154	5069_60917
… 74 more positive …
… 49917 more negative …

Weight^?	Feature
+0.480	9937_16385
+0.428	9937_16396
+0.380	9937_16396 9937_16385
+0.365	9937_16385 9937_16396
+0.301	9937_16396 9937_16385 9937_16396
+0.295	9937_16385 9937_16396 9937_16385
+0.138	9937_46614
+0.101	9937_46594
+0.076	9937_16385 9937_16385
+0.074	9716_62468
… 460 more positive …
… 49531 more negative …

Weight^?	Feature
+0.343	9934_34305
+0.342	9716_19972
+0.266	9052_49924
+0.219	9934_58115
+0.211	9934_62977
+0.187	5078_26557
+0.182	9716_4098
+0.180	9716_19969
+0.172	9716_4118
+0.169	9716_19980
… 384 more positive …
… 49607 more negative …

Weight^?	Feature
+0.501	9916_25100
+0.382	9916_25089 9916_25100
+0.377	9916_25089
+0.370	9916_25100 9916_25089
+0.267	9916_25089 9916_25100 9916_25089
+0.243	9916_25100 9916_25089 9916_25100
+0.110	9916_25100 9916_25100
+0.108	9916_25100 9916_25089 9916_25089
+0.100	9916_25089 9916_25089 9916_25100
+0.097	9916_25089 9916_25089
… 754 more positive …
… 49237 more negative …

Weight^?	Feature
+0.434	7755_853
+0.395	9752_15475
+0.250	9091_55767
+0.233	7755_38846
+0.178	9710_42229
+0.178	7755_18823
+0.152	7755_30238
+0.150	9752_14633
+0.137	5061_966
+0.125	9091_55762
… 143 more positive …
… 49848 more negative …

Weight^?	Feature
+0.434	9701_64774
+0.285	7700_6193
+0.283	9701_64800
+0.265	9701_64771
+0.242	9739_24836
+0.217	7716_3672
+0.174	7700_6193 9701_64774
+0.161	9701_64774 7700_6193
+0.145	9701_14085
+0.145	7700_6192
… 448 more positive …
… 49543 more negative …

Weight^?	Feature
+0.575	9739_21270
+0.288	9739_21250 9739_21270
+0.278	9739_21250
+0.266	9739_21270 9739_21250
+0.237	9739_21270 9739_21250 9739_21270
+0.221	9739_21250 9739_21270 9739_21250
+0.200	7755_15810
+0.183	9739_21270 7755_15810
+0.183	9739_21270 9739_21270
+0.179	7755_15810 9739_21270
… 776 more positive …
… 49215 more negative …

Weight^?	Feature
+0.406	9716_61953
+0.317	9716_16897
+0.304	9916_42755
+0.291	9716_61964
+0.261	9716_61956
+0.229	7716_62856
+0.189	9716_33025
+0.166	7716_28965
+0.141	9716_1824
+0.127	9716_61953 9716_16897
… 485 more positive …
… 49506 more negative …

Weight^?	Feature
+0.315	9947_534
+0.277	9947_514
+0.266	9741_9987
+0.242	9741_10016
+0.224	9708_29186
+0.218	9708_29206
+0.214	9741_9986
+0.197	9741_10006
+0.177	9741_10016 9741_9987
+0.164	9708_29206 9708_29186
… 655 more positive …
… 49336 more negative …

Weight^?	Feature
+0.397	9758_43523
+0.257	9053_45315
+0.225	9916_65282
+0.215	9758_55574
+0.207	9916_65302
+0.175	9716_39425
+0.168	9053_57632
+0.157	9735_13068
+0.157	7752_1337
+0.139	7752_1336
… 716 more positive …
… 49275 more negative …

Weight^?	Feature
+0.549	9935_44310
+0.459	9935_44290
+0.285	9935_44310 9935_44290
+0.277	9935_44290 9935_44310
+0.273	9935_44292
+0.200	9935_44290 9935_44310 9935_44290
+0.197	9935_44310 9935_44290 9935_44310
+0.100	9935_44310 9935_44292
+0.095	9935_8449
+0.091	9935_44310 9935_44310
… 740 more positive …
… 49251 more negative …

Weight^?	Feature
+0.294	9937_29185
+0.276	9732_50437
+0.223	9937_29196
+0.222	9716_27424
+0.212	9716_19974
+0.209	9716_27395
+0.209	9732_5378
+0.198	9705_2049
+0.186	9716_19971
+0.178	9937_28161
… 823 more positive …
… 49168 more negative …

Weight^?	Feature
+0.468	7716_3579
+0.397	9716_34606
+0.263	7716_3579 9716_34606
+0.253	9716_34606 7716_3579
+0.204	9745_10307
+0.149	9745_10314
+0.137	9716_9218
+0.137	9754_12964
+0.131	7716_3579 9716_34606 7716_3579
+0.126	9716_34609
… 137 more positive …
… 49854 more negative …

Weight^?	Feature
+0.371	9732_27168
+0.371	9732_27139
+0.335	9732_27168 9732_27139
+0.286	9732_27139 9732_27168
+0.250	9934_16130
+0.215	9732_27168 9732_27139 9732_27168
+0.215	9934_16150
+0.213	9732_27139 9732_27168 9732_27139
+0.194	9934_36609
+0.184	9934_16150 9934_16130
… 270 more positive …
… 49721 more negative …

Weight^?	Feature
+0.626	7755_6981
+0.501	7755_6981 7755_6981
+0.340	7755_6981 7755_6981 7755_6981
+0.243	9745_10316
+0.202	7755_62987
+0.146	9745_10303 9745_10316
+0.125	7749_25397
+0.121	9754_44238
+0.109	9745_10303
+0.091	9745_10308 9754_44238
… 23 more positive …
… 49968 more negative …

Weight^?	Feature
+0.507	7757_21260
+0.487	7743_21261
+0.330	7743_21261 7757_21260
+0.324	7757_21260 7743_21261
+0.228	7757_21260 7757_21260
+0.183	7757_21260 7743_21261 7757_21260
+0.183	7743_21261 7757_21260 7743_21261
+0.177	7743_21261 7743_21261
+0.146	9757_21261
+0.138	7743_21261 7757_21260 7757_21260
… 54 more positive …
… 49937 more negative …

Weight^?	Feature
+0.546	9735_26627
+0.380	9937_1537
+0.283	9736_11780
+0.264	9736_11782
+0.219	9735_19462
+0.204	9716_55558
+0.152	9735_6150
+0.149	7752_61571
+0.145	9935_47873
+0.141	9736_33025
… 496 more positive …
… 49495 more negative …

Weight^?	Feature
+0.490	7755_32046
+0.376	7755_35296
+0.275	9752_15575
+0.237	7755_32046 7755_35296
+0.227	9767_11599
+0.226	7755_30238
+0.160	7755_38846
+0.157	7755_35296 7755_32046
+0.151	7755_35299
+0.137	7755_38846 7755_32046
… 161 more positive …
… 49830 more negative …

Weight^?	Feature
+0.447	7755_26527
+0.275	9752_14267
+0.249	7755_26527 9700_32785
+0.249	7755_26527 9700_32785 9752_14633
+0.240	9700_32785 9752_14633
+0.192	7755_31336
+0.190	9700_32785
+0.171	9716_7033
+0.167	7755_26527 7755_26527
+0.161	9752_14633
… 173 more positive …
… 49818 more negative …

Weight^?	Feature
+0.472	9916_21516
+0.428	9916_21505
+0.297	9916_21505 9916_21516
+0.286	9916_21516 9916_21505
+0.240	9734_48387
+0.180	9916_21516 9916_21505 9916_21516
+0.175	9916_21505 9916_21516 9916_21505
+0.122	9734_48385
+0.105	9916_21536
+0.095	9774_54788
… 846 more positive …
… 49145 more negative …

Weight^?	Feature
+0.471	9934_56066
+0.372	9934_35329
+0.328	9934_56086
+0.242	9934_35340
+0.207	9934_56068
+0.206	7752_3098
+0.194	9934_56066 9934_35329
+0.188	9934_35329 9934_56066
+0.183	7752_52368
+0.150	9934_35332
… 461 more positive …
… 49530 more negative …

Weight^?	Feature
+0.366	9934_24834
+0.364	9934_24854
+0.305	5043_5890
+0.263	5043_50966
+0.260	9934_24834 9934_24854
+0.260	9934_24854 9934_24834
+0.229	9934_24833
+0.191	5043_5890 5043_50966
+0.187	5043_50966 5043_5890
+0.157	9934_24854 9934_24834 9934_24854
… 665 more positive …
… 49326 more negative …

Weight^?	Feature
+0.246	9788_34562
+0.231	9734_14337
+0.220	9934_6657
+0.210	9717_12802
+0.207	9717_12805
+0.182	9734_14341
+0.179	9711_4353
+0.150	9717_4613
+0.146	9934_51714
+0.143	9717_60673
… 899 more positive …
… 49092 more negative …

Weight^?	Feature
+0.411	9916_20226
+0.394	9916_20246
+0.327	9916_20226 9916_20246
+0.321	9916_20246 9916_20226
+0.250	9916_20226 9916_20246 9916_20226
+0.250	9916_20246 9916_20226 9916_20246
+0.215	9739_29953
+0.209	9739_10502
+0.203	9739_10502 9739_29953
+0.197	9739_29953 9739_10502
… 593 more positive …
… 49398 more negative …

Weight^?	Feature
+0.356	9716_2049
+0.347	9716_2052
+0.344	9716_2060
+0.212	5005_65285
+0.203	9935_32534
+0.191	9716_2060 9716_2049
+0.184	9716_2049 9716_2060
+0.180	9935_11523
+0.177	5005_8736
+0.152	9716_2052 9716_2049
… 567 more positive …
… 49424 more negative …

Weight^?	Feature
+0.511	7716_2147
+0.490	7716_38835
+0.211	7716_2866
+0.202	7716_65523
+0.195	7716_62858
+0.185	7716_36730
+0.185	7716_6026
+0.166	7716_38835 7716_38835
+0.141	9716_36734
+0.133	9716_33838
… 99 more positive …
… 49892 more negative …

Weight^?	Feature
+0.463	9758_57092
+0.328	7747_3968
+0.323	9734_12803
+0.282	9734_12832
+0.168	9935_32514
+0.165	9734_12803 9758_57092
+0.158	7702_4341
+0.145	9758_57092 9734_12803
+0.140	7746_3617
+0.136	9758_57092 9734_12832
… 706 more positive …
… 49285 more negative …

Weight^?	Feature
+0.463	7755_33266
+0.418	7755_55462
+0.229	7755_33266 7755_55462
+0.206	7755_6988
+0.178	7755_55462 7755_33266
+0.149	7700_13970
+0.140	7752_6193
+0.131	7754_35304
+0.119	7700_37286
+0.111	7700_803 7700_13970
… 284 more positive …
… 49707 more negative …

Weight^?	Feature
+0.472	9935_62467
+0.372	9935_62496
+0.341	9935_62466
+0.332	9935_62486
+0.211	9935_62467 9935_62466
+0.193	9935_62466 9935_62467
+0.174	9935_62486 9935_62467
+0.166	9935_62466 9935_62486
+0.165	9935_62496 9935_62486
+0.155	9935_62486 9935_62496
… 602 more positive …
… 49389 more negative …

Weight^?	Feature
+0.375	7700_6192
+0.367	7716_3497 7716_3497
+0.366	7716_3497 7716_3497 7716_3497
+0.312	7716_18164
+0.312	7716_3497
+0.296	9700_14516
+0.226	9700_8631
+0.213	7716_18165
+0.183	7700_6192 9700_14516
+0.176	7716_18164 7716_18164
… 16 more positive …
… 49975 more negative …

Weight^?	Feature
+0.463	9916_23809
+0.340	9935_53516
+0.274	9916_3333
+0.235	9734_12812 9734_12801
+0.234	9916_3340
+0.223	9734_12801 9734_12812
+0.219	9734_12801 9734_12812 9734_12801
+0.213	9734_12812
+0.211	9734_12801
+0.203	9734_12812 9734_12801 9734_12812
… 540 more positive …
… 49451 more negative …

Weight^?	Feature
+0.333	9916_42764
+0.310	7755_15800
+0.295	7755_6982
+0.285	9916_42753
+0.235	9754_51222
+0.234	9754_51202
+0.196	7755_15801
+0.190	7752_1959
+0.150	9754_43526
+0.139	7752_1941
… 747 more positive …
… 49244 more negative …

Weight^?	Feature
+0.419	9716_62466
+0.360	9716_62486
+0.333	9716_62466 9716_62486
+0.330	9716_62486 9716_62466
+0.271	9716_62466 9716_62486 9716_62466
+0.257	9716_62469
+0.247	9716_62486 9716_62466 9716_62486
+0.158	9716_62469 9716_62466
+0.155	9716_62466 9716_62469
+0.140	7797_40544
… 657 more positive …
… 49334 more negative …

Weight^?	Feature
+0.444	7752_3103
+0.246	9752_33921
+0.246	9734_33312 9734_33283
+0.234	9734_33312
+0.232	9734_33283 9734_33312
+0.229	9734_33312 9734_33283 9734_33312
+0.221	9734_33283
+0.215	9734_33283 9734_33312 9734_33283
+0.205	9934_49921
+0.182	9752_3090
… 650 more positive …
… 49341 more negative …

Weight^?	Feature
+0.341	9934_46624
+0.324	9739_27905 9739_27905
+0.303	9934_46595
+0.303	9740_8972
+0.284	9739_27905 9739_27905 9739_27905
+0.273	9739_27905
+0.216	9916_62476 9916_62465 9916_62476
+0.194	9916_62476
+0.190	9916_62476 9916_62465
+0.178	9916_62465 9916_62476
… 38 more positive …
… 49953 more negative …

Weight^?	Feature
+0.505	9937_42241
+0.422	9937_35585
+0.367	9937_42252
+0.347	9937_35596
+0.203	9937_35585 9937_42241
+0.177	9937_42241 9937_35585
+0.169	9937_35596 9937_42252
+0.146	9937_42252 9937_35596
+0.138	9937_42244
+0.136	7757_8040
… 478 more positive …
… 49513 more negative …

Weight^?	Feature
+0.448	7716_35363
+0.275	7716_25270
+0.250	7716_3976
+0.236	7716_35365
+0.222	9716_20252
+0.216	9716_20251
+0.211	7716_35364
+0.193	7752_7903
+0.179	7752_7749
+0.178	9716_33844
… 178 more positive …
… 49813 more negative …

Weight^?	Feature
+0.375	7746_6555
+0.364	7752_59117
+0.287	7752_59118
+0.248	9953_32901
+0.243	7716_25354
+0.241	7752_59117 7752_59118
+0.234	7752_59118 7752_59117
+0.194	7752_59117 7752_59118 7752_59117
+0.174	9953_32901 7746_6555
+0.168	7716_35360
… 419 more positive …
… 49572 more negative …

Weight^?	Feature
+0.548	7752_4705
+0.301	7752_56488
+0.285	5069_57369
+0.222	5015_57368
+0.190	7752_45995
+0.190	5069_57364
+0.187	7752_36321
+0.148	7755_36135
+0.143	7752_4705 7752_56488
+0.143	7755_62891
… 244 more positive …
… 49747 more negative …

Weight^?	Feature
+0.385	9734_30486
+0.378	9734_47107
+0.351	9734_47136
+0.325	9734_30466
+0.202	9734_30466 9734_47107
+0.179	9734_47107 9734_30466
+0.166	9734_30486 9734_30466
+0.162	9734_47136 9734_30486
+0.160	9734_30466 9734_30486
+0.157	9734_47107 9734_47136
… 280 more positive …
… 49711 more negative …

Weight^?	Feature
+0.412	7716_33417
+0.394	9716_33096
+0.262	7716_33417 9716_33096
+0.252	9716_33096 7716_33417
+0.246	9736_48406
+0.215	9734_5635
+0.214	9736_48386
+0.197	9736_48406 9736_48386
+0.193	9736_48386 9736_48406
+0.155	9736_48386 9736_48406 9736_48386
… 611 more positive …
… 49380 more negative …

Weight^?	Feature
+0.507	9916_46338
+0.306	9739_44320
+0.257	9739_44291
+0.221	9916_46339
+0.201	9916_46338 9916_46338
+0.196	9739_44320 9739_44291
+0.187	9916_46338 9916_20993
+0.185	9739_44291 9739_44320
+0.180	9916_20993 9916_46338
+0.154	9916_20993
… 917 more positive …
… 49074 more negative …

Weight^?	Feature
+0.424	7742_36320
+0.393	7742_5774
+0.320	7700_3394
+0.266	7727_52509
+0.246	7716_5458
+0.227	7700_3391
+0.226	7716_60012
+0.176	7742_5774 7742_36320
+0.172	7716_5458 7700_3394
+0.149	7700_3394 7716_5458
… 218 more positive …
… 49773 more negative …

Weight^?	Feature
+0.368	9734_61728
+0.352	9734_61699 9734_61728
+0.351	9734_61699
+0.320	9734_61728 9734_61699
+0.276	9734_61728 9734_61699 9734_61728
+0.272	9734_61699 9734_61728 9734_61699
+0.203	9934_56067
+0.201	9934_56065
+0.199	9934_49666
+0.109	9934_56076
… 776 more positive …
… 49215 more negative …

Weight^?	Feature
+0.424	9916_771 9916_800
+0.373	9916_800 9916_771
+0.346	9916_771
+0.341	9916_771 9916_800 9916_771
+0.335	9916_800 9916_771 9916_800
+0.319	9916_800
+0.254	7755_37288
+0.181	9739_56580
+0.115	9739_56580 7755_37288
+0.105	7755_37288 9739_56580
… 467 more positive …
… 49524 more negative …

Weight^?	Feature
+0.503	9000_6150
+0.466	9000_6147
+0.243	9000_56066
+0.224	9734_64258
+0.221	9000_6150 9000_6147
+0.219	9000_6147 9000_6150
+0.214	9734_64278
+0.175	9734_64258 9734_64278
+0.172	9734_64278 9734_64258
+0.148	9000_6150 9000_6150
… 357 more positive …
… 49634 more negative …

Weight^?	Feature
+0.480	9916_5121
+0.405	9916_5132
+0.298	7752_3570
+0.241	9734_64258
+0.203	9734_64278
+0.171	9734_64258 9734_64278
+0.168	9916_5132 9916_5121
+0.168	9916_5121 9916_5132
+0.153	9734_64278 9734_64258
+0.141	9734_64258 9734_64278 9734_64258
… 581 more positive …
… 49410 more negative …

Weight^?	Feature
+0.429	7752_3100
+0.396	9752_14478
+0.293	7700_33790
+0.229	7752_3100 9752_14478
+0.229	9752_14478 7752_3100
+0.224	7752_6193
+0.170	7700_33784
+0.151	7752_5712
+0.142	7752_51351
+0.131	7752_30043
… 261 more positive …
… 49730 more negative …

Weight^?	Feature
+0.510	9734_61728
+0.381	9734_61699
+0.269	9734_5633
+0.224	9734_61699 9734_61728
+0.213	9734_61702
+0.207	9734_61728 9734_61699
+0.188	9734_5633 9734_61728
+0.178	9734_61728 9734_5633
+0.152	9734_61728 9734_61728
+0.149	9734_5644
… 245 more positive …
… 49746 more negative …

Weight^?	Feature
+0.371	7716_3323
+0.356	7752_1981
+0.278	9752_15548
+0.261	7752_1981 9752_15548
+0.258	9752_15548 7752_1981
+0.229	7716_3323 9716_33229
+0.214	9716_33229
+0.201	9716_33229 7716_3323
+0.185	9716_33102 7716_3323
+0.144	9716_33102
… 360 more positive …
… 49631 more negative …

Weight^?	Feature
+0.406	9716_33107
+0.318	9716_6427
+0.310	9752_509
+0.227	7730_3507 9716_6427
+0.202	7730_3803
+0.197	9752_509 9700_8601
+0.197	9716_33107 7730_3803
+0.197	9716_6427 9716_33107
+0.176	7730_3803 7755_38849
+0.159	9716_33107 9716_33107
… 101 more positive …
… 49890 more negative …

Weight^?	Feature
+0.347	9934_12038
+0.256	7752_1350
+0.251	9739_52741
+0.246	7752_1346
+0.172	9770_10154
+0.164	9733_39685
+0.153	9916_3586
+0.150	9754_43526
+0.150	7755_33269
+0.147	9752_1333
… 706 more positive …
… 49285 more negative …

Weight^?	Feature
+0.622	7716_21024
+0.403	9716_20729
+0.287	9716_20649
+0.268	9716_20729 7716_21024
+0.221	7716_21024 9716_20729
+0.148	7716_21024 9716_20649
+0.145	9716_20649 7716_21024
+0.122	9716_17280
+0.118	7716_21024 9716_20729 7716_21024
+0.108	7716_21024 7716_21024
… 357 more positive …
… 49634 more negative …

Weight^?	Feature
+0.475	9776_50694
+0.339	9776_50691
+0.322	9776_53765
+0.231	9043_28421
+0.208	9043_28420
+0.196	9734_47110
+0.183	9776_50720
+0.151	9776_50694 9776_50691
+0.146	9776_53765 9776_50694
+0.142	9776_50694 9776_53765
… 489 more positive …
… 49502 more negative …

Weight^?	Feature
+0.288	9716_6427
+0.278	9716_20257
+0.271	9716_20257 7730_3803
+0.248	7730_3803
+0.247	9752_19080
+0.211	9716_6427 9716_20257
+0.173	9716_6427 9716_6427
+0.161	7730_3803 7755_38849
+0.151	7755_38849
+0.151	9716_6427 9716_6427 9716_20257
… 124 more positive …
… 49867 more negative …

Weight^?	Feature
+0.336	9734_5634 9734_5654
+0.333	9734_5634
+0.309	9734_5654 9734_5634
+0.291	9734_5654
+0.291	9734_5634 9734_5654 9734_5634
+0.269	9916_65283
+0.268	9916_65312 9916_65283
+0.258	9916_65312
+0.256	9734_5654 9734_5634 9734_5654
+0.250	9916_65283 9916_65312
… 830 more positive …
… 49161 more negative …

Weight^?	Feature
+0.508	7716_3387
+0.330	7755_15818
+0.261	7716_3386
+0.256	7755_15810
+0.204	7755_15818 7716_3387
+0.182	7755_15810 7716_3387
+0.174	7716_3387 7755_15818
+0.172	7716_3385
+0.148	7716_3387 7755_15810
+0.147	7755_15830
… 322 more positive …
… 49669 more negative …

Weight^?	Feature
+0.437	9935_27916
+0.408	9935_27905
+0.365	9739_24326
+0.226	9716_7174
+0.174	9935_27916 9935_27905
+0.171	9935_27905 9935_27916
+0.167	9716_7171
+0.157	9739_24326 9716_7171
+0.141	9716_48388
+0.138	9716_7171 9739_24326
… 658 more positive …
… 49333 more negative …

Weight^?	Feature
+0.750	7757_2510
+0.653	7757_2510 7757_2510
… 49991 more negative …
-0.006	9702_30576
-0.006	9700_8607
-0.006	9916_20226
-0.006	9734_61702
-0.007	9702_30571
-0.008	9752_14642
-0.009	9752_14633
-0.009	<BIAS>

Weight^?	Feature
+0.229	9773_4612
+0.195	9071_19715
+0.172	9773_4610
+0.168	9739_10499
+0.165	9759_52993
+0.139	9739_10528
+0.133	9773_4357
+0.129	7765_6723
+0.126	9754_51232
+0.123	9754_51203
… 887 more positive …
… 49104 more negative …

Weight^?	Feature
+0.512	5046_53704
+0.512	9091_57626
+0.327	9091_57626 5046_53704
+0.326	5046_53704 9091_57626
+0.218	5046_53704 5046_53704
+0.206	9091_57626 9091_57626
+0.186	5046_53704 9091_57626 5046_53704
+0.179	9091_57626 5046_53704 9091_57626
+0.146	9091_57626 5046_53704 5046_53704
+0.146	5046_53704 5046_53704 9091_57626
… 189 more positive …
… 49802 more negative …

Weight^?	Feature
+0.353	9713_41730
+0.294	9713_2562
+0.282	9716_30210 9716_30230 9716_30210
+0.273	9716_30230 9716_30210
+0.267	9716_30230 9716_30210 9716_30230
+0.253	9900_64002
+0.246	9716_30210 9716_30230
+0.213	9900_64001
+0.202	9716_30210
+0.202	9713_41730 9713_2562
… 184 more positive …
… 49807 more negative …

Weight^?	Feature
+0.472	9739_44290
+0.413	9739_44310
+0.312	9739_44290 9739_44310
+0.303	9739_44310 9739_44290
+0.298	9934_33537
+0.207	9739_44310 9739_44290 9739_44310
+0.203	9739_44290 9739_44310 9739_44290
+0.179	7752_3385
+0.138	9739_44293
+0.119	7752_3385 9934_33537
… 559 more positive …
… 49432 more negative …

Weight^?	Feature
+0.683	9953_32909
+0.364	9746_37923
+0.306	9953_32909 9746_37923
+0.243	9746_37923 9953_32909
+0.191	9953_32909 9746_37923 9953_32909
+0.174	9746_37926
+0.139	9746_37927
+0.139	9746_37926 9953_32909
+0.139	9953_32909 9746_37926
+0.121	9953_32909 9746_37927
… 106 more positive …
… 49885 more negative …

Weight^?	Feature
+0.450	9752_13162
+0.217	9091_52619
+0.216	7752_13178
+0.205	7752_13176
+0.204	9752_13168
+0.188	9752_13162 7752_13176
+0.173	9091_52614
+0.173	9011_56717
+0.169	9752_19471
+0.164	9716_3665
… 247 more positive …
… 49744 more negative …

Weight^?	Feature
+0.423	9716_13058
+0.274	9716_9219
+0.239	9916_35852
+0.211	9916_35841
+0.210	9716_48899
+0.202	9716_25366
+0.194	9916_35852 9916_35841
+0.189	9916_35841 9916_35852
+0.165	9716_3841
+0.155	9716_50689
… 733 more positive …
… 49258 more negative …

Weight^?	Feature
+0.278	9732_39180
+0.246	9732_39169
+0.239	9732_49932
+0.200	9916_46339
+0.196	9732_49921
+0.171	7767_25554
+0.147	9732_40192
+0.137	9732_49921 9732_49932
+0.122	9900_64513
+0.122	9736_55043
… 1040 more positive …
… 48951 more negative …

Weight^?	Feature
+0.548	9716_31787
+0.348	9746_16333 9716_31787
+0.341	9746_16333
+0.292	9716_31787 9746_16333
+0.266	9716_31787 9746_16333 9716_31787
+0.219	9716_31787 9746_16330
+0.207	9746_16330
+0.197	9746_16330 9716_31787
+0.178	9716_31787 9746_16330 9716_31787
+0.157	9746_16333 9716_31787 9746_16333
… 173 more positive …
… 49818 more negative …

Weight^?	Feature
+0.499	7752_12914
+0.336	7746_6554
+0.269	9752_31489 7752_12914
+0.249	9752_31489
+0.215	7752_12914 9752_31489
+0.198	9752_30414
+0.171	9752_12950
+0.161	7752_12914 9752_31489 7752_12914
+0.145	7752_30411
+0.140	7746_6553
… 327 more positive …
… 49664 more negative …

Weight^?	Feature
+0.625	9752_10853 9752_10853
+0.560	9752_10853
+0.533	9752_10853 9752_10853 9752_10853
+0.015	<BIAS>
… 49991 more negative …
-0.006	9700_8607
-0.006	9916_20226
-0.006	9734_61702
-0.007	9702_30571
-0.008	9752_14642
-0.009	9752_14633

Weight^?	Feature
+0.507	9716_32506
+0.279	9716_33092 9716_32506
+0.263	9716_33092
+0.246	9716_32506 9716_33092
+0.188	9716_32506 9716_33092 9716_32506
+0.176	7798_14544
+0.175	9767_1125
+0.175	9716_33096 9716_32506
+0.151	9716_32506 9716_32506
+0.147	9716_33096
… 150 more positive …
… 49841 more negative …

Weight^?	Feature
+0.476	7716_3976
+0.410	7716_3976 9716_33102
+0.403	9716_33102 7716_3976
+0.380	9716_33102
+0.229	7716_3976 9716_33102 7716_3976
+0.211	9716_33102 7716_3976 9716_33102
+0.193	7716_3976 7716_3976
+0.174	9716_33102 7716_3976 7716_3976
+0.171	7716_3976 7716_3976 9716_33102
+0.152	9716_33102 9716_33102 7716_3976
… 200 more positive …
… 49791 more negative …

Weight^?	Feature
+0.337	9701_36610
+0.277	9900_14594
+0.276	9716_13827
+0.242	9701_16130
+0.224	9701_36609
+0.211	9701_16129
+0.192	9716_58882
+0.179	9716_58902
+0.174	9934_39425
+0.140	9701_16133
… 808 more positive …
… 49183 more negative …

Weight^?	Feature
+0.412	7716_48445
+0.278	7716_62858
+0.226	7716_2147
+0.206	7700_14753
+0.200	9700_32786
+0.189	7752_44414
+0.158	7716_38835
+0.157	7752_44413
+0.157	7716_48445 7716_62858
+0.148	7716_6026
… 169 more positive …
… 49822 more negative …

Weight^?	Feature
+0.597	9716_33109
+0.398	7784_36266
+0.320	9716_33109 9716_33109
+0.285	7784_36266 7784_36266
+0.228	9716_33109 9700_33819
+0.174	9757_9090
+0.173	9700_33819
+0.141	9716_33104
+0.123	9752_1938
+0.108	9716_33107
… 48 more positive …
… 49943 more negative …

Weight^?	Feature
+0.546	9083_40964
+0.420	9083_40961
+0.210	9083_40961 9083_40964
+0.199	9083_40964 9083_40961
+0.170	9716_19974
+0.168	9083_40964 9083_40964
+0.168	9083_40972
+0.156	9716_27424
+0.152	9716_27395
+0.126	9083_53505
… 371 more positive …
… 49620 more negative …

Weight^?	Feature
+0.489	9736_11782
+0.336	9716_48386
+0.303	9716_48406
+0.265	9935_63748
+0.166	9716_48389
+0.135	9716_18178
+0.133	9735_24324
+0.129	9716_48406 9716_48386
+0.129	9736_11782 9736_11782
+0.125	9735_24322
… 824 more positive …
… 49167 more negative …

Weight^?	Feature
+0.460	9734_64257
+0.402	9734_64268
+0.395	9734_64268 9734_64257
+0.362	9734_64257 9734_64268
+0.325	9734_64257 9734_64268 9734_64257
+0.302	9734_64268 9734_64257 9734_64268
+0.211	9734_64260
+0.111	7716_902
+0.079	9734_64257 9734_64260
+0.070	9734_64260 9734_64257
… 368 more positive …
… 49623 more negative …

Weight^?	Feature
+0.417	9732_39180
+0.374	9732_39172
+0.247	9732_39172 9732_39180
+0.242	9732_39180 7755_9519
+0.230	7755_35295 9732_39172
+0.220	9732_39169
+0.193	9739_47361
+0.179	7755_35295 9732_39172 9732_39180
+0.179	7755_9519 9732_39172
+0.159	7755_9519
… 53 more positive …
… 49938 more negative …

Weight^?	Feature
+0.358	9752_14479
+0.355	7716_35365
+0.239	7752_3257
+0.207	9752_14479 7752_3257
+0.181	7752_3257 9752_14479
+0.179	9752_1949
+0.172	9752_1948
+0.155	7716_35364
+0.149	7752_21240
+0.133	7716_35364 7716_35365
… 323 more positive …
… 49668 more negative …

Weight^?	Feature
+0.462	9752_8589
+0.354	9752_8586
+0.301	9702_30571
+0.251	9752_8586 9702_30571
+0.249	9752_8589 9702_30571
+0.228	9702_30571 9752_8586
+0.204	9702_30571 9752_8589
+0.199	9702_30576
+0.195	9702_30576 9752_8589
+0.177	9752_8589 9702_30576
… 217 more positive …
… 49774 more negative …

Weight^?	Feature
+0.528	9700_36237
+0.439	9700_36237 7700_8608
+0.403	7700_8608 9700_36237
+0.298	7700_8608
+0.292	9700_36237 7700_8608 9700_36237
+0.250	7700_8608 9700_36237 7700_8608
+0.139	9700_36237 9700_36237
+0.100	9700_36237 7700_1152
+0.099	7700_1152 9700_36237
+0.098	7700_8608 7700_8608
… 116 more positive …
… 49875 more negative …

Weight^?	Feature
+0.561	7725_30004
+0.328	9752_15544
+0.200	7752_46836
+0.172	7755_15811
+0.162	7725_30004 7755_15803
+0.143	7755_15832
+0.139	9702_30571
+0.138	7752_45684
+0.137	7755_15803
+0.137	7755_15832 7725_30004
… 531 more positive …
… 49460 more negative …

Weight^?	Feature
+0.588	7752_3917
+0.394	9752_19201
+0.331	9752_19205
+0.257	9752_19205 7752_3917
+0.246	7752_3917 9752_19201
+0.232	9752_19201 7752_3917
+0.207	7752_3917 9752_19205
+0.157	7752_3917 9752_19205 7752_3917
+0.153	7752_3918
+0.135	7752_3917 9752_19201 7752_3917
… 206 more positive …
… 49785 more negative …

Weight^?	Feature
+0.317	7755_30229
+0.276	7755_30228
+0.268	7755_30224
+0.256	9752_10376
+0.252	5061_550 7755_30228
+0.252	5061_1340 9752_10376
+0.245	5061_1340
+0.197	9752_2305 5061_550
+0.181	5061_550
+0.168	9752_10376 9752_2305 5061_550
… 72 more positive …
… 49919 more negative …

Weight^?	Feature
+0.435	9754_51232
+0.400	9754_51203
+0.369	9754_51232 9754_51203
+0.366	9754_51203 9754_51232
+0.319	9754_51203 9754_51232 9754_51203
+0.316	9754_51232 9754_51203 9754_51232
+0.159	9916_55809
+0.151	9739_10500
+0.104	7752_13272
+0.084	9754_51232 9754_51232
… 535 more positive …
… 49456 more negative …

Weight^?	Feature
+0.300	7716_3936
+0.298	7716_60041
+0.236	9900_55812
+0.214	9716_39389
+0.213	9716_39388
+0.201	9736_43778
+0.198	7700_801
+0.193	7745_46877
+0.181	9716_50689
+0.177	9716_6407
… 610 more positive …
… 49381 more negative …

Weight^?	Feature
+0.469	7755_31330
+0.288	7757_7013
+0.225	5061_590
+0.213	7755_31330 7755_31330
+0.193	5061_595
+0.160	5061_550
+0.133	7757_7041
+0.132	7755_37277
+0.131	9716_31062
+0.124	7755_31330 7716_3669
… 290 more positive …
… 49701 more negative …

Weight^?	Feature
+0.487	9734_14851
+0.475	9734_14880
+0.342	9734_14880 9734_14851
+0.327	9734_14851 9734_14880
+0.241	9734_14880 9734_14851 9734_14880
+0.228	9734_14851 9734_14880 9734_14851
+0.178	9935_14851
+0.170	9758_23300
+0.103	9734_14851 9734_14851
+0.100	9734_14880 9734_14880
… 449 more positive …
… 49542 more negative …

Weight^?	Feature
+0.423	9716_2049
+0.393	9716_2052
+0.355	9739_22530
+0.348	9716_2060
+0.167	9716_2060 9716_2049
+0.156	9716_2049 9716_2060
+0.151	9716_2052 9739_22530
+0.150	9716_2052 9716_2049
+0.140	9739_22530 9716_2052
+0.123	9716_2049 9716_2052
… 416 more positive …
… 49575 more negative …

Weight^?	Feature
+0.413	7716_3161
+0.364	7716_49282
+0.328	9716_3151
+0.300	9734_5382
+0.267	9935_32514
+0.198	9935_32534
+0.193	9716_31052
+0.172	9935_32514 9935_32534
+0.157	9935_32534 9935_32514
+0.121	9935_32517
… 588 more positive …
… 49403 more negative …

Weight^?	Feature
+0.591	7716_3669
+0.230	7752_26210
+0.221	7716_3669 7716_3669
+0.215	7755_307
+0.202	7755_44698
+0.175	7755_32044
+0.169	7755_2984
+0.167	7755_307 7716_3669
+0.147	7755_2987
+0.135	7716_3669 7755_307
… 316 more positive …
… 49675 more negative …

Weight^?	Feature
+0.430	9736_11782
+0.362	9935_63748
+0.310	9735_13088
+0.264	9735_13059
+0.245	9736_11782 9935_63748
+0.243	9935_63748 9736_11782
+0.210	9735_13088 9735_13059
+0.210	9735_13059 9735_13088
+0.157	9935_63748 9736_11782 9935_63748
+0.154	9735_13059 9735_13088 9735_13059
… 574 more positive …
… 49417 more negative …

Weight^?	Feature
+0.481	9915_7682
+0.324	9937_28172
+0.267	9937_28161
+0.197	9937_35585
+0.179	9790_45569
+0.179	5032_515
+0.150	9915_10754
+0.141	9915_10774
+0.141	9915_7681
+0.120	7743_21745
… 748 more positive …
… 49243 more negative …

Weight^?	Feature
+0.335	7700_55154
+0.328	9754_51222
+0.315	9754_51202
+0.213	9754_51202 9754_51222
+0.195	9754_18690
+0.188	9754_51222 9754_51202
+0.185	7700_5871
+0.175	9701_49154
+0.172	9946_21250
+0.155	9754_51222 9754_51202 9754_51222
… 886 more positive …
… 49105 more negative …

Weight^?	Feature
+0.385	7755_33348
+0.337	7755_35297
+0.280	9742_467
+0.199	9787_13993
+0.199	9743_1232
+0.191	7755_15802
+0.183	9742_21018
+0.169	7755_6805
+0.144	5061_770
+0.134	9716_3722
… 143 more positive …
… 49848 more negative …

Weight^?	Feature
+0.310	7746_6159
+0.293	7746_31256
+0.280	7752_515
+0.266	7746_33171
+0.251	7752_516
+0.220	9746_38504
+0.195	7716_9597
+0.162	7730_3433
+0.155	7746_6551
+0.146	7716_62603
… 251 more positive …
… 49740 more negative …

Weight^?	Feature
+0.466	9953_6547
+0.371	9953_6545
+0.272	5061_1788 5061_1788
+0.272	9953_6547 9953_6545
+0.232	5061_1788 5061_1788 5061_1788
+0.218	9752_40873
+0.207	9953_6547 9953_6547
+0.199	9953_6545 9953_6545
+0.182	5061_1788
+0.174	9752_10376 9752_40873
… 33 more positive …
… 49958 more negative …

Weight^?	Feature
+0.519	9716_31786
+0.204	9716_6427
+0.166	7730_3431
+0.158	9702_30571
+0.153	9716_44960
+0.149	9716_31786 7730_3431
+0.145	7730_3507 9716_6427 9716_6427
+0.143	9752_8588
+0.139	9716_31786 9716_31786
+0.133	9716_6427 9716_6427
… 147 more positive …
… 49844 more negative …

Weight^?	Feature
+0.447	9700_730
+0.339	9700_49395
+0.274	9752_30042
+0.271	9716_37507
+0.266	9754_3559
+0.218	9700_14259
+0.194	9754_26704
+0.176	9716_37507 7730_380
+0.159	9700_14256
+0.136	9700_730 9752_30042
… 106 more positive …
… 49885 more negative …

Weight^?	Feature
+0.444	7752_6313
+0.337	7752_6313 9752_14642
+0.309	9752_14642
+0.296	9752_14642 7752_6313
+0.283	7752_6313 7752_6313
+0.256	7752_6313 9752_14642 7752_6313
+0.216	7752_6313 7752_6313 9752_14642
+0.206	9752_14642 7752_6313 7752_6313
+0.200	9752_14642 7752_6313 9752_14642
+0.187	9752_14642 9752_14642
… 321 more positive …
… 49670 more negative …

Weight^?	Feature
+0.385	9935_27424
+0.341	9935_27395
+0.285	9935_41740
+0.217	9935_27395 9935_27424
+0.193	5002_17410
+0.191	9935_27424 9935_27395
+0.191	9935_41729
+0.165	9935_47875
+0.162	9900_24854
+0.156	9935_27395 9935_27424 9935_27395
… 570 more positive …
… 49421 more negative …

Weight^?	Feature
+0.480	9716_33838
+0.409	7716_6166
+0.250	7755_31331
+0.250	7755_48439
+0.234	9716_33838 7716_6166
+0.213	7716_6166 9716_33838
+0.195	9716_33838 9716_33838
+0.122	7716_6166 7716_6166
+0.114	7716_6166 9716_33838 9716_33838
+0.114	9716_33838 9716_33838 7716_6166
… 146 more positive …
… 49845 more negative …

Weight^?	Feature
+0.591	7755_33346
+0.215	7755_33346 7755_33346
+0.195	9716_3722
+0.186	9752_30046
+0.175	9700_32785
+0.164	7755_33346 9700_32785
+0.156	7755_14939
+0.154	9778_58579
+0.141	7784_30136
+0.123	7750_8494
… 190 more positive …
… 49801 more negative …

Weight^?	Feature
+0.619	7700_1611
+0.314	7700_1632
+0.297	7752_30043
+0.219	7700_1633
+0.207	7752_30043 7700_1611
+0.186	7716_2267
+0.181	7752_3100
+0.150	7700_1611 7752_30043
+0.126	9716_2279
+0.107	7700_1611 7700_1611
… 423 more positive …
… 49568 more negative …

Weight^?	Feature
+0.398	9937_48652
+0.368	9937_48641
+0.310	9937_48652 9937_48641
+0.310	9937_48641 9937_48652
+0.252	9937_48641 9937_48652 9937_48641
+0.236	9937_48652 9937_48641 9937_48652
+0.227	9937_26912
+0.196	9937_44576
+0.136	9937_44576 9937_26912
+0.124	9937_26912 9937_44576
… 643 more positive …
… 49348 more negative …

Weight^?	Feature
+0.499	9716_30220
+0.444	9716_30209
+0.276	9716_30209 9716_30220
+0.273	9716_30220 9716_30209
+0.187	9716_30209 9716_30220 9716_30209
+0.180	9716_30220 9716_30209 9716_30220
+0.149	9716_30230
+0.143	9916_25100
+0.137	9916_25100 9916_25089
+0.135	9916_25089 9916_25100
… 446 more positive …
… 49545 more negative …

Weight^?	Feature
+0.362	9953_37016
+0.250	9752_34079
+0.237	9716_34606 7730_3431 5061_910
+0.218	7730_3431 5061_910
+0.218	9716_34606
+0.211	7730_3431 5061_910 9752_10379
+0.188	5061_910 9752_10379
+0.184	9752_10379 9752_34079
+0.181	9716_34606 7730_3431
+0.181	5061_550 7730_3047
… 117 more positive …
… 49874 more negative …

Weight^?	Feature
+0.374	9745_10301
+0.371	9745_12668
+0.319	7745_12684
+0.312	7755_44696
+0.236	7745_59010
+0.236	7755_32044
+0.188	7745_8223
+0.182	7755_44696 7755_32044
+0.173	7745_12684 9745_12668
+0.171	9745_10301 7745_59010
… 226 more positive …
… 49765 more negative …

Weight^?	Feature
+0.299	9774_41987
+0.281	9790_33281
+0.267	9734_30466
+0.259	9774_10243
+0.193	7742_479
+0.180	5022_19796
+0.168	9774_10243 9774_41987
+0.161	9790_33281 9790_33281
+0.157	5022_19796 5022_19796
+0.148	9734_47105
… 353 more positive …
… 49638 more negative …

Weight^?	Feature
+0.665	9934_32513
+0.276	9934_12038
+0.252	9934_12037
+0.236	7752_36575
+0.198	9934_12054
+0.197	9934_12038 9934_32513
+0.187	9934_32513 9934_12038
+0.178	7752_5986
+0.123	9934_32513 9934_32513
+0.116	9934_12034
… 587 more positive …
… 49404 more negative …

Weight^?	Feature
+0.394	9716_7173
+0.253	9739_29956
+0.230	9789_48918
+0.180	9789_46595
+0.178	9716_13057
+0.170	9715_40452
+0.168	9916_55810
+0.167	9789_48898
+0.166	9716_15106
+0.165	9716_7173 9739_29956
… 543 more positive …
… 49448 more negative …

Weight^?	Feature
+0.352	9739_44293
+0.316	9701_47366
+0.313	9754_40452
+0.244	9701_47876
+0.221	9705_33028
+0.191	9739_44293 9739_44293
+0.182	9701_25861
+0.163	9934_30466
+0.148	9701_47366 9701_25861
+0.147	7755_38450
… 182 more positive …
… 49809 more negative …

Weight^?	Feature
+0.561	9937_48641
+0.389	9937_36865
+0.337	9937_29187
+0.216	9937_29188
+0.169	9716_39941
+0.168	9937_1538
+0.164	9937_48641 9937_36865
+0.153	9937_36865 9937_48641
+0.145	9937_48641 9937_48641
+0.127	9937_29187 9937_48641
… 326 more positive …
… 49665 more negative …

Weight^?	Feature
+0.486	7755_33266
+0.398	9716_6428
+0.306	7755_33266 9716_6428
+0.227	9716_6428 7755_33266
+0.198	9752_1375
+0.162	9716_6427 7755_33266
+0.152	7755_33266 9716_6428 7755_33266
+0.143	9716_6428 7730_3803
+0.142	9716_6428 7730_3803 5061_517
+0.141	7755_33269
… 98 more positive …
… 49893 more negative …

Weight^?	Feature
+0.425	7742_62681
+0.359	9914_29187
+0.242	7742_6020
+0.234	9734_64258
+0.209	9734_64278
+0.209	9790_60930
+0.195	9734_64261
+0.170	9734_64278 9734_64258
+0.156	9734_48396
+0.153	9734_64258 9734_64278
… 672 more positive …
… 49319 more negative …

Weight^?	Feature
+0.368	7716_12111
+0.280	7752_20141
+0.230	7716_7045
+0.206	7752_42111
+0.202	7716_28965
+0.188	7752_2087
+0.176	7752_6751
+0.166	9752_10374
+0.161	7716_7044
+0.154	5078_59092
… 321 more positive …
… 49670 more negative …

Weight^?	Feature
+0.295	9734_5633
+0.291	9734_5644
+0.281	9734_5633 9734_5644
+0.267	9734_5644 9734_5633
+0.254	9734_5633 9734_5644 9734_5633
+0.242	9734_5644 9734_5633 9734_5644
+0.227	9734_61702
+0.214	9916_65283
+0.210	9916_65312
+0.210	9916_65283 9916_65312
… 824 more positive …
… 49167 more negative …

Weight^?	Feature
+0.413	9935_55041
+0.411	9935_55052
+0.265	9739_56582
+0.264	9935_55052 9935_55041
+0.251	9739_56608
+0.243	9739_56579
+0.241	9935_55041 9935_55052
+0.189	9935_55041 9935_55052 9935_55041
+0.169	9935_55052 9935_55041 9935_55052
+0.151	9739_56579 9739_56608
… 650 more positive …
… 49341 more negative …

Weight^?	Feature
+0.497	9716_13059
+0.416	9716_25376
+0.296	9716_25376 9716_13059
+0.265	9716_13059 9716_25376
+0.195	9937_37635
+0.193	9716_13059 9716_13059
+0.187	9716_25347
+0.161	9716_25376 9716_25376
+0.153	7757_46774
+0.143	9716_25376 9716_13059 9716_25376
… 780 more positive …
… 49211 more negative …

Weight^?	Feature
+0.646	7716_62344
+0.333	9716_33391
+0.280	9716_30014
+0.213	7716_62344 7716_62344
+0.189	7716_33390
+0.179	7716_62344 9716_33391
+0.165	9716_33391 7716_62344
+0.165	9716_14854
+0.139	9734_8197
+0.138	7716_62344 9716_30014
… 434 more positive …
… 49557 more negative …

Weight^?	Feature
+0.431	9736_56096
+0.424	9736_56067
+0.361	9736_56067 9736_56096
+0.303	9736_56096 9736_56067
+0.211	9736_56096 9736_56067 9736_56096
+0.197	9736_56067 9736_56096 9736_56067
+0.146	9736_56067 9736_56067
+0.137	9716_62467 9716_62496
+0.136	9736_27137
+0.135	9736_56096 9736_56096
… 731 more positive …
… 49260 more negative …

Weight^?	Feature
+0.485	9935_48897
+0.470	9935_48908
+0.357	9935_48908 9935_48897
+0.315	9935_48897 9935_48908
+0.224	9935_48908 9935_48897 9935_48908
+0.219	9935_48897 9935_48908 9935_48897
+0.147	9739_10508
+0.144	9739_10500
+0.139	9935_48897 9935_48897
+0.105	9935_48908 9935_48908
… 807 more positive …
… 49184 more negative …

Weight^?	Feature
+0.431	7752_61571
+0.322	9735_7172
+0.235	9734_30468
+0.217	9734_30486
+0.198	7752_61571 9735_7172
+0.184	9735_7172 7752_61571
+0.182	9736_36374
+0.172	9734_47110
+0.157	9735_7169
+0.150	9734_30466
… 780 more positive …
… 49211 more negative …

Weight^?	Feature
+0.340	9734_33284
+0.291	9734_5378
+0.270	9735_7169
+0.245	9935_63748
+0.242	9734_5381
+0.227	7752_61575
+0.217	9716_10497
+0.176	9735_7172
+0.172	9734_5398
+0.168	9716_50689
… 709 more positive …
… 49282 more negative …

Weight^?	Feature
+0.439	9732_39200
+0.357	9732_39171
+0.343	9732_27139
+0.228	9732_27139 9732_39200
+0.219	9732_27168
+0.208	9732_39200 9732_27139
+0.189	9732_39171 9732_39200
+0.182	9732_39200 9732_39171
+0.178	9732_39200 9732_27168
+0.144	9732_39171 9732_39200 9732_39171
… 522 more positive …
… 49469 more negative …

Weight^?	Feature
+0.418	7716_31983
+0.411	7755_3818
+0.379	7716_36161
+0.361	7755_36118
+0.205	7755_3818 7716_31983
+0.180	7716_31983 7755_3818
+0.159	7755_36118 7716_36161
+0.159	7716_36161 7755_36118
+0.155	7755_36118 7716_31983
+0.153	7716_36161 7755_3818
… 285 more positive …
… 49706 more negative …

Weight^?	Feature
+0.543	9716_13827
+0.404	9716_43779
+0.279	7716_6166
+0.263	9716_43779 9716_13827
+0.230	9716_13827 9716_43779
+0.175	9716_64516
+0.165	9716_13827 9716_13827
+0.149	9716_64517
+0.132	9716_43779 9716_13827 9716_43779
+0.120	9716_43779 9716_43779
… 410 more positive …
… 49581 more negative …

Weight^?	Feature
+0.357	9752_14478
+0.272	9716_6427
+0.232	5061_1195 7730_2658
+0.227	5061_1195
+0.224	5061_1195 7730_2658 9716_6427
+0.221	7730_3803 5061_517
+0.217	7730_3803 5061_517 9752_14478
+0.216	7730_2658 9716_6427
+0.203	5061_517 9752_14478
+0.183	9752_14478 5061_1195
… 52 more positive …
… 49939 more negative …

Weight^?	Feature
+0.433	9937_46851
+0.365	9716_13058
+0.364	7757_1047
+0.324	7743_1045
+0.230	9937_28162
+0.162	9757_1045
+0.148	9937_46851 7757_1047
+0.144	9716_13058 9716_13058
+0.123	9937_46851 7743_1045
+0.110	9937_28162 9937_46851
… 605 more positive …
… 49386 more negative …

Weight^?	Feature
+0.412	9716_35306
+0.361	9716_33096
+0.253	5060_17339
+0.208	9752_10379
+0.193	9700_33819
+0.190	9752_10379 9700_1655
+0.190	9752_10379 9700_1655 9716_35306
+0.190	9716_33096 9716_33096
+0.179	9752_30046
+0.178	9716_35306 9716_33096
… 61 more positive …
… 49930 more negative …

Weight^?	Feature
+0.545	9916_59906
+0.257	9901_4354
+0.221	9916_59906 9916_33025
+0.216	9901_26881
+0.208	9916_33025 9916_59906
+0.182	9751_19713
+0.182	9916_59906 9916_59906
+0.162	9751_19716
+0.131	9916_33025
+0.130	9751_19713 9751_19716
… 797 more positive …
… 49194 more negative …

Weight^?	Feature
+0.615	9953_6546
+0.289	7746_6554
+0.269	7746_6555
+0.258	7746_6552
+0.213	9953_6547
+0.208	7746_6554 9953_6546
+0.195	9953_6546 7746_6552
+0.170	7746_6552 9953_6546
+0.162	9953_6546 7746_6555
+0.145	9953_6546 7746_6554
… 358 more positive …
… 49633 more negative …

Weight^?	Feature
+0.532	9716_19970
+0.451	9716_19990
+0.322	9716_19970 9716_19990
+0.320	9716_19990 9716_19970
+0.225	9716_19970 9716_19990 9716_19970
+0.222	9716_19973
+0.221	9716_19990 9716_19970 9716_19990
+0.146	9716_19970 9716_19970
+0.137	9716_19973 9716_19970
+0.132	9716_19970 9716_19973
… 303 more positive …
… 49688 more negative …

Weight^?	Feature
+0.554	7755_6981
+0.374	9700_1653
+0.310	7755_6980
+0.258	7755_6980 7755_6981
+0.203	7755_6981 7755_6980
+0.164	7755_6981 7755_6981
+0.157	5077_58279
+0.157	9700_1653 9700_1635
+0.143	9700_1635
+0.129	7755_6981 7755_6980 7755_6981
… 121 more positive …
… 49870 more negative …

Weight^?	Feature
+0.347	9916_1793
+0.345	9734_5637
+0.323	9916_1804
+0.316	9916_1793 9916_1804
+0.303	9916_1804 9916_1793
+0.280	9916_1793 9916_1804 9916_1793
+0.269	9916_1804 9916_1793 9916_1804
+0.226	9734_5634
+0.207	9734_5637 9734_5634
+0.199	9734_5634 9734_5637
… 426 more positive …
… 49565 more negative …

Weight^?	Feature
+0.378	7752_3235
+0.331	9752_2904
+0.330	7749_15177
+0.307	7752_3235 9752_2904
+0.289	9752_2904 7752_3235
+0.269	7729_31441
+0.191	7752_3235 9752_2904 7752_3235
+0.178	7749_15177 7729_31441
+0.176	7729_31441 7749_15177
+0.171	7752_3235 7752_3235
… 273 more positive …
… 49718 more negative …

Weight^?	Feature
+0.427	7757_3276
+0.425	9757_46770
+0.369	7757_1053
+0.239	9757_46770 7757_3276
+0.225	7757_3276 9757_46770
+0.193	7743_46770
+0.138	5061_595
+0.126	7757_1053 9757_46770
+0.126	7716_2219
+0.122	7746_6551
… 240 more positive …
… 49751 more negative …

Weight^?	Feature
+0.500	9934_12033
+0.396	7752_8145
+0.241	7752_5986
+0.207	9934_12044
+0.203	7752_8145 9934_12033
+0.190	9934_12033 7752_8145
+0.178	9716_14853
+0.167	9734_59905
+0.159	9934_12033 9934_12033
+0.121	9934_15872
… 804 more positive …
… 49187 more negative …

Weight^?	Feature
+0.505	9734_61702
+0.366	9734_16642
+0.320	9734_16642 9734_61702
+0.301	9734_61702 9734_16642
+0.169	9734_16642 9734_61702 9734_16642
+0.167	9734_61702 9734_61702
+0.165	9734_61699
+0.162	9934_61186
+0.158	9734_61702 9734_16642 9734_61702
+0.139	9734_61699 9734_61702
… 685 more positive …
… 49306 more negative …

Weight^?	Feature
+0.469	9916_41475
+0.461	9916_41504
+0.350	9916_41504 9916_41475
+0.344	9916_41475 9916_41504
+0.269	9916_41475 9916_41504 9916_41475
+0.258	9916_41504 9916_41475 9916_41504
+0.152	9732_39180
+0.137	9732_39169
+0.103	9916_41475 9916_41475
+0.103	9732_40196
… 661 more positive …
… 49330 more negative …

Weight^?	Feature
+0.506	9937_48641
+0.497	9937_48652
+0.298	9937_42242
+0.228	9937_42262
+0.224	9937_48652 9937_48641
+0.217	9937_48641 9937_48652
+0.146	9937_48641 9937_42242
+0.121	9937_48641 9937_48652 9937_48641
+0.119	9937_42262 9937_48652
+0.119	9937_42242 9937_48641
… 583 more positive …
… 49408 more negative …

Weight^?	Feature
+0.331	7772_101
+0.328	5061_1721
+0.316	7716_3251
+0.314	9734_33284
+0.217	7772_34134
+0.203	7772_101 5061_1721
+0.169	9705_26627
+0.161	7772_34134 5061_1721
+0.147	9734_47105
+0.138	9716_32507
… 225 more positive …
… 49766 more negative …

Weight^?	Feature
+0.416	7745_3090
+0.380	9752_13212
+0.273	7752_7816
+0.270	7745_3091
+0.255	9752_13217
+0.190	7752_3290
+0.159	7746_33183
+0.157	7752_61571
+0.155	7716_55676
+0.152	9745_3082
… 433 more positive …
… 49558 more negative …

Weight^?	Feature
+0.422	7716_65296
+0.399	7746_6509
+0.378	7716_65295
+0.372	7746_62518
+0.225	7746_6509 7716_65296
+0.216	7716_65296 7746_6509
+0.181	7716_65295 7746_62518
+0.177	7746_62518 7716_65295
+0.117	7746_6509 7716_65295
+0.104	7746_62518 7716_65296
… 313 more positive …
… 49678 more negative …

Weight^?	Feature
+0.506	9752_8585
+0.311	9709_3656
+0.284	9709_3657
+0.234	9702_30571
+0.214	9702_30576
+0.213	9752_8585 9702_30571
+0.183	9702_30571 9752_8585
+0.181	9752_46818
+0.165	9752_8585 9702_30576
+0.132	9709_3656 9752_8585
… 303 more positive …
… 49688 more negative …

Weight^?	Feature
+0.692	7757_2262
+0.350	9757_10810
+0.280	9937_38150
+0.214	7757_2262 9757_10810
+0.206	9757_10810 7757_2262
+0.191	7757_2262 7757_2262
+0.162	9937_38150 7757_2262
+0.162	7757_2262 9937_38150
+0.115	7757_2262 9757_10810 7757_2262
+0.112	9757_10810 7757_2262 9757_10810
… 648 more positive …
… 49343 more negative …

Weight^?	Feature
+0.292	9937_37644
+0.280	9734_9996
+0.272	9734_30465
+0.271	9937_37633
+0.253	9734_61702
+0.213	9734_16641
+0.210	9937_37633 9937_37644
+0.184	9734_9985
+0.182	9937_37644 9937_37633
+0.167	9734_61728
… 859 more positive …
… 49132 more negative …

Weight^?	Feature
+0.571	9916_21506
+0.471	9916_21526
+0.262	7752_3591
+0.197	9752_19471
+0.159	7752_19479
+0.156	7752_3591 9916_21506
+0.146	7752_3592
+0.134	9745_2096
+0.130	9916_21506 9916_21526
+0.129	9916_21536
… 553 more positive …
… 49438 more negative …

Weight^?	Feature
+0.444	9752_13217
+0.420	9745_11838
+0.320	9745_11838 9752_13217
+0.287	9745_11838 9752_13217 9745_11838
+0.268	9752_13217 9745_11838
+0.251	9752_13217 9745_11838 9752_13217
+0.215	9752_13217 9752_13217 9752_13217
+0.209	9752_13217 9752_13217
+0.197	9716_11890
+0.185	9745_11838 9752_13212
… 37 more positive …
… 49954 more negative …

Weight^?	Feature
+0.328	9916_62469
+0.325	9916_51458
+0.300	9916_25089
+0.254	9916_62486
+0.252	9916_62466
+0.232	9916_32770
+0.199	9720_12293
+0.170	9916_62466 9916_62486
+0.154	9916_62486 9916_62466
+0.124	9794_31489
… 555 more positive …
… 49436 more negative …

Weight^?	Feature
+0.516	9701_2818
+0.379	9752_30042
+0.255	9701_2818 9752_30042
+0.222	9701_2817
+0.213	9752_30042 9701_2818
+0.209	9739_24324
+0.191	7700_815
+0.179	9701_2817 9701_2818
+0.146	9700_33783
+0.132	9701_2818 9701_2817
… 603 more positive …
… 49388 more negative …

Weight^?	Feature
+0.381	9734_48386
+0.306	9734_48406
+0.271	7716_44974
+0.219	9734_3078
+0.211	9734_48406 9734_48386
+0.210	9934_45313
+0.206	9934_45313 9934_45324
+0.205	9934_45313 9934_45324 9934_45313
+0.204	9734_48386 9734_48406
+0.201	9734_48388
… 609 more positive …
… 49382 more negative …

Weight^?	Feature
+0.372	9916_20226
+0.347	9916_20246
+0.312	9758_34564
+0.295	9916_20226 9916_20246
+0.294	9916_20246 9916_20226
+0.246	9916_20246 9916_20226 9916_20246
+0.241	9916_20226 9916_20246 9916_20226
+0.239	9746_34226
+0.135	7746_14896
+0.111	9746_14880
… 854 more positive …
… 49137 more negative …

Weight^?	Feature
+0.414	9734_48406
+0.412	9734_48386
+0.362	9734_48386 9734_48406
+0.330	9734_48406 9734_48386
+0.283	9734_48406 9734_48386 9734_48406
+0.266	9734_48386 9734_48406 9734_48386
+0.158	9715_40452
+0.125	7752_7412
+0.112	9734_48388
+0.083	9734_48406 9734_48388
… 324 more positive …
… 49667 more negative …

Weight^?	Feature
+0.428	9746_14891
+0.207	9746_3610
+0.207	9754_10278 5061_498
+0.194	9716_33827
+0.181	5061_1720 5061_1721 9700_10918
+0.180	9746_14891 9746_14891
+0.176	5061_498
+0.172	5061_1720
+0.170	9746_14893
+0.162	5061_1721 9700_10918
… 92 more positive …
… 49899 more negative …

Weight^?	Feature
+0.481	7716_3944
+0.341	7755_15809
+0.329	7755_3817
+0.248	7755_55462
+0.202	7755_15809 7716_3944
+0.191	7716_3944 7755_15809
+0.191	7716_3944 7755_3817
+0.180	7716_3937
+0.152	7755_6988
+0.150	7745_59010
… 217 more positive …
… 49774 more negative …

Weight^?	Feature
+0.227	9700_14256 9754_26709
+0.223	9716_6408
+0.217	9700_14256
+0.213	9754_26709
+0.213	9700_49395 9700_14256
+0.202	9700_49395
+0.202	9700_14256 9754_26709 5061_498
+0.202	9700_49395 9700_14256 9754_26709
+0.202	9754_26709 5061_498
+0.195	9743_8542
… 101 more positive …
… 49890 more negative …

Weight^?	Feature
+0.565	7799_39909
+0.417	7757_65159
+0.267	7799_39960
+0.240	7757_65154
+0.185	7757_65159 7799_39909
+0.175	7799_39909 7757_65159
+0.147	5007_5119
+0.147	7797_63090
+0.134	7799_39909 7757_65154
+0.134	7757_65154 7799_39909
… 239 more positive …
… 49752 more negative …

Weight^?	Feature
+0.475	9701_47362
+0.337	9701_47382
+0.252	9701_47365
+0.249	9701_25859
+0.247	9701_25861
+0.238	7716_25270
+0.203	9701_25888
+0.154	9701_47382 9701_47362
+0.149	9716_62469
+0.144	9716_9219
… 537 more positive …
… 49454 more negative …

Weight^?	Feature
+0.492	9739_56582
+0.366	9934_56067
+0.246	9739_56608
+0.228	9934_56076
+0.204	9739_22529
+0.204	9739_56579
+0.189	9739_10502
+0.181	9934_11009
+0.173	9739_24837
+0.165	9934_56065
… 539 more positive …
… 49452 more negative …

Weight^?	Feature
+0.586	9716_26298
+0.339	9716_26298 7730_3431
+0.226	7730_3431
+0.212	9716_26298 9716_26298
+0.182	9716_26295
+0.179	9700_32786
+0.154	7755_2985
+0.141	9716_26298 9716_26298 7730_3431
+0.122	9700_32786 7755_2985
+0.121	9716_34606
… 192 more positive …
… 49799 more negative …

Weight^?	Feature
+0.540	7755_2456
+0.449	7755_35903
+0.358	7755_2456 7755_35903
+0.353	7755_35903 7755_2456
+0.259	7755_2456 7755_35903 7755_2456
+0.237	7755_35903 7755_2456 7755_35903
+0.142	7755_33288
+0.093	9757_12117
+0.088	9757_12058
+0.083	7755_2456 7755_2456
… 112 more positive …
… 49879 more negative …

Weight^?	Feature
+0.571	7752_37850
+0.358	9752_11191
+0.278	9752_19491
+0.258	7752_19490
+0.241	7752_37850 7752_37850
+0.196	7752_37850 9752_11191
+0.196	9752_10378
+0.174	9752_11191 7752_37850
+0.155	9752_19491 7752_37850
+0.151	7752_37850 9752_19491
… 227 more positive …
… 49764 more negative …

Weight^?	Feature
+0.332	9716_61955
+0.320	9716_61984
+0.290	9748_2
+0.254	9744_63235
+0.240	9716_61955 9716_61984
+0.217	9716_61984 9716_61955
+0.203	9716_61958
+0.163	9716_61984 9716_61955 9716_61984
+0.160	9716_16898
+0.156	9934_30467
… 798 more positive …
… 49193 more negative …

Weight^?	Feature
+0.527	9716_34609
+0.284	9752_864
+0.257	9716_33229
+0.203	9716_34609 9716_33229
+0.201	7730_3431
+0.196	9716_34609 9716_34609
+0.191	7755_38849
+0.185	9716_34609 7730_3431
+0.142	9716_34609 9716_34609 7730_3431
+0.142	9752_14633 9752_864
… 152 more positive …
… 49839 more negative …

Weight^?	Feature
+0.558	9934_6659
+0.336	9934_45313 9934_6659
+0.325	9934_6659 9934_45313
+0.324	9934_45313
+0.210	9934_6659 9934_6659
+0.169	9934_6659 9934_45324
+0.168	9934_45324
+0.164	9934_6659 9934_45313 9934_6659
+0.144	9934_45324 9934_6659
+0.133	9934_6659 9934_6659 9934_45313
… 605 more positive …
… 49386 more negative …

Weight^?	Feature
+0.614	9935_14850
+0.384	7752_19204
+0.285	9752_10504
+0.217	9935_14850 9935_14850
+0.208	9934_40963
+0.207	7752_19204 9935_14850
+0.201	9935_14850 7752_19204
+0.191	7752_3291
+0.133	9752_10504 9935_14850
+0.118	9935_14850 9752_10504
… 407 more positive …
… 49584 more negative …

Weight^?	Feature
+0.302	9709_3657
+0.272	9709_3655
+0.189	7755_30226
+0.183	9709_3657 9709_3655
+0.182	9700_43994
+0.177	7731_1545
+0.173	5061_966
+0.171	9752_1879
+0.169	9752_46818
+0.166	9754_40450
… 219 more positive …
… 49772 more negative …

Weight^?	Feature
+0.355	9716_13058
+0.320	7716_60325
+0.300	9716_25346
+0.210	9916_55810
+0.194	9716_25366
+0.191	9752_15946
+0.161	9716_25346 7716_60325
+0.158	7752_15942
+0.138	7716_60325 9716_13058
+0.134	9716_9217
… 693 more positive …
… 49298 more negative …

Weight^?	Feature
+0.286	7755_30226
+0.268	9709_3657
+0.227	9752_46818
+0.218	9752_8585
+0.216	5061_1340
+0.210	7752_1880
+0.200	7755_30228
+0.187	7755_30226 5061_1340
+0.167	5009_55924
+0.157	9709_3654
… 196 more positive …
… 49795 more negative …

Weight^?	Feature
+0.464	9716_32508
+0.273	9716_32508 7730_3431
+0.248	9716_32508 7730_3431 5061_910
+0.244	9716_32508 9716_32508
+0.233	9757_2490
+0.226	7730_3431 5061_910
+0.223	7730_3431 5061_910 9757_46770
+0.223	5061_910 9757_46770
+0.213	9757_1045
+0.198	9716_32508 9716_32508 7730_3431
… 49 more positive …
… 49942 more negative …

Weight^?	Feature
+0.333	7755_30226
+0.276	9752_1879
+0.217	7755_30228
+0.213	7755_30226 5061_1340
+0.204	5061_1427
+0.200	9752_46818
+0.196	7755_30228 7755_30226
+0.186	5061_1340
+0.166	7755_37877
+0.147	5061_1427 7755_30228
… 202 more positive …
… 49789 more negative …

Weight^?	Feature
+0.540	9007_54871
+0.508	9016_51473
+0.311	9007_54871 9016_51473
+0.311	9016_51473 9007_54871
+0.209	9016_51473 9007_54871 9016_51473
+0.203	9007_54871 9016_51473 9007_54871
+0.198	9016_51478
+0.112	9007_54871 9016_51478
+0.109	9016_51478 9007_54871
+0.091	9701_24835
… 469 more positive …
… 49522 more negative …

Weight^?	Feature
+0.372	9746_16333
+0.341	9701_21475
+0.266	9701_21475 9701_21475
+0.228	9701_21475 9701_21475 9701_21475
+0.204	9716_34609
+0.183	9746_16334
+0.173	9953_6545
+0.159	5061_3615
+0.152	9746_16330
+0.135	9953_39154
… 159 more positive …
… 49832 more negative …

Weight^?	Feature
+0.379	9916_41475
+0.289	9916_41504
+0.286	9734_14870
+0.255	9916_33025
+0.237	9934_36609
+0.204	9916_33036
+0.198	9734_14850
+0.197	9734_14850 9734_14870
+0.181	7752_14263
+0.172	9734_14870 9734_14850
… 718 more positive …
… 49273 more negative …

Weight^?	Feature
+0.540	9716_34606
+0.274	7716_60325
+0.228	9916_25110
+0.214	9716_34606 9716_34606
+0.199	7752_8746
+0.196	9716_34606 7716_60325
+0.195	7716_3578
+0.189	7716_60325 9716_34606
+0.170	9916_25090
+0.168	9716_9217
… 523 more positive …
… 49468 more negative …

Weight^?	Feature
+0.440	9935_53516
+0.391	9935_53505
+0.272	9935_53505 9935_53516
+0.270	9739_24844 9739_24833
+0.264	9739_24844
+0.257	9739_24833
+0.248	9739_24833 9739_24844
+0.248	9935_53516 9935_53505
+0.212	9739_24844 9739_24833 9739_24844
+0.207	9739_24833 9739_24844 9739_24833
… 509 more positive …
… 49482 more negative …

Weight^?	Feature
+0.331	7755_6984
+0.314	9902_60674
+0.244	9739_10497
+0.232	9739_10500
+0.183	9902_40214
+0.165	9902_40194
+0.149	5061_550
+0.146	7767_17635
+0.146	5061_55
+0.142	7755_6936
… 366 more positive …
… 49625 more negative …

Weight^?	Feature
+0.462	7700_55154
+0.381	7700_5871
+0.263	9716_62468
+0.233	9705_46338
+0.176	9701_49157
+0.174	9716_62476
+0.173	9701_42245
+0.155	9716_28684
+0.151	9701_42242
+0.142	9705_262
… 574 more positive …
… 49417 more negative …

Weight^?	Feature
+0.430	9734_48388
+0.318	9734_48387 9734_48388
+0.301	9734_48387
+0.286	9734_48388 9734_48387
+0.277	9934_24834
+0.216	9934_24854
+0.196	9734_48388 9734_48387 9734_48388
+0.193	9934_45313
+0.178	7752_12914
+0.158	9734_48388 9734_48388
… 625 more positive …
… 49366 more negative …

Weight^?	Feature
+0.471	9916_19458
+0.452	9752_14043
+0.265	7752_21101
+0.210	7752_23980
+0.190	9752_14043 9916_19458
+0.185	9916_19458 9752_14043
+0.181	9734_33286
+0.180	7716_3253
+0.151	9752_14043 7752_21101
+0.134	7752_21101 9752_14043
… 388 more positive …
… 49603 more negative …

Weight^?	Feature
+0.435	7752_21240
+0.421	7752_21240 7752_21240
+0.360	7752_21240 7752_21240 7752_21240
+0.264	7755_31332
+0.225	7752_1959
+0.173	7716_3668
+0.170	9752_15548
+0.169	7755_31332 7716_3669 7755_31332
+0.161	7702_15765
+0.155	7716_3669 7755_31332
… 58 more positive …
… 49933 more negative …

Weight^?	Feature
+0.435	9716_25346
+0.364	7716_25849
+0.296	9716_13058
+0.271	9716_25366
+0.205	9916_16130
+0.178	9916_16150
+0.175	9716_28135
+0.169	7716_25983
+0.137	9716_25346 7716_25849
+0.136	9716_25346 9716_25346
… 716 more positive …
… 49275 more negative …

Weight^?	Feature
+0.459	7752_892
+0.405	9916_16130
+0.313	9916_16150
+0.269	9916_16132
+0.202	7716_60049
+0.160	9734_5637
+0.138	9734_5634
+0.135	9916_16130 7752_892
+0.133	9916_16130 9916_16150
+0.133	9916_16150 9916_16130
… 632 more positive …
… 49359 more negative …

Weight^?	Feature
+0.433	7716_31050
+0.406	9702_30733
+0.285	7702_6616
+0.270	7702_65328
+0.203	7702_374
+0.188	9708_32621
+0.142	7716_59660
+0.141	9716_31052
+0.130	9702_30733 7702_6616
+0.130	7702_6616 9702_30733
… 415 more positive …
… 49576 more negative …

Weight^?	Feature
+0.365	9935_63750
+0.331	9937_59906
+0.316	9739_5377
+0.296	9739_5388
+0.282	7752_61571
+0.205	9739_5398
+0.202	9739_24325
+0.178	9739_5378
+0.161	9739_5377 9739_5388
+0.155	9739_5388 9739_5377
… 490 more positive …
… 49501 more negative …

Weight^?	Feature
+0.407	9937_48652
+0.353	9937_48641
+0.313	9937_48641 9937_48652
+0.301	9937_48652 9937_48641
+0.225	9937_48652 9937_48641 9937_48652
+0.225	9937_48641 9937_48652 9937_48641
+0.200	9742_20492
+0.173	9742_20481
+0.153	9903_37377
+0.137	9902_42242
… 731 more positive …
… 49260 more negative …

Weight^?	Feature
+0.394	9916_26390
+0.366	9916_26370
+0.264	9916_26370 9916_26390
+0.240	9916_26390 9916_26370
+0.195	9736_8449
+0.192	9916_26390 9916_26370 9916_26390
+0.176	9720_63748
+0.160	9720_63745
+0.145	9720_63756
+0.144	9736_8460
… 933 more positive …
… 49058 more negative …

Weight^?	Feature
+0.587	9716_33106
+0.333	7716_1831
+0.258	7716_1831 9716_33106
+0.256	9716_33106 7716_1831
+0.185	7716_1843
+0.184	9716_33106 9716_33106
+0.166	7716_3976
+0.158	9716_33106 7716_1831 9716_33106
+0.148	9716_33106 7716_3976
+0.129	9716_1832
… 241 more positive …
… 49750 more negative …

Weight^?	Feature
+0.591	9716_62011
+0.290	7716_55795
+0.287	7716_55654
+0.271	7716_55654 9716_62011
+0.252	9716_62011 7716_55654
+0.241	7716_55795 9716_62011
+0.223	9716_62011 7716_55795
+0.183	9716_62011 7716_55654 9716_62011
+0.178	9716_62011 9716_62011
+0.157	7716_55654 9716_62011 7716_55654
… 203 more positive …
… 49788 more negative …

Weight^?	Feature
+0.535	9716_31787
+0.354	9716_31787 9746_16333
+0.351	9746_16333 9716_31787
+0.339	9746_16333
+0.306	9716_31787 9746_16333 9716_31787
+0.224	9746_16333 9716_31787 9746_16333
+0.203	9746_16330 9716_31787
+0.180	9746_16330
+0.157	9716_31787 9746_16330
+0.145	9716_31787 9746_16330 9716_31787
… 161 more positive …
… 49830 more negative …

Weight^?	Feature
+0.321	9734_12812 9734_12801
+0.304	9734_12812
+0.301	9734_12801 9734_12812
+0.293	9734_12801
+0.277	9734_12801 9734_12812 9734_12801
+0.268	9734_12812 9734_12801 9734_12812
+0.243	9937_40193
+0.235	9937_19724
+0.227	9937_28161
+0.197	9937_19716
… 431 more positive …
… 49560 more negative …

Weight^?	Feature
+0.439	9705_46337
+0.205	9701_23298
+0.201	9734_5637
+0.187	9908_45078
+0.171	9701_23299
+0.164	9734_10006
+0.157	9908_45058
+0.154	9908_45078 9908_45058
+0.153	9734_9986
+0.152	9903_49153
… 777 more positive …
… 49214 more negative …

Weight^?	Feature
+0.536	7741_5926
+0.409	9741_48901
+0.275	9900_6659
+0.266	9900_10753
+0.182	9741_48901 7741_5926
+0.180	7741_5926 9741_48901
+0.179	9014_60422
+0.158	7741_5926 7741_5926
+0.141	9900_6668
+0.137	5028_4824
… 506 more positive …
… 49485 more negative …

Weight^?	Feature
+0.489	9709_3657
+0.329	9752_8585
+0.320	9752_46818
+0.266	9709_3656
+0.227	9709_3657 9752_46818
+0.227	9752_46818 9709_3657
+0.191	9709_3657 9752_8585
+0.153	9709_3657 9752_46818 9709_3657
+0.139	9752_46818 9709_3657 9752_46818
+0.127	9752_8585 9709_3657
… 251 more positive …
… 49740 more negative …

Weight^?	Feature
+0.629	7755_30226
+0.495	7755_30226 7755_30226
+0.388	7755_30226 7755_30226 7755_30226
+0.208	7755_30229 7755_30226 7755_30226
+0.191	7755_30229 7755_30226
+0.182	7755_30229
+0.182	7755_30226 7755_30226 7755_30229
+0.161	7755_30226 7755_30229
+0.130	7755_30226 7755_30229 7755_30226
+0.042	7755_30229 7755_30229 7755_30226
… 1 more positive …
… 49990 more negative …

Weight^?	Feature
+0.449	7716_6166
+0.413	9716_33838
+0.315	9716_33838 7716_6166
+0.312	7716_6166 9716_33838
+0.184	9716_33838 7716_6166 9716_33838
+0.167	7716_6166 7716_6166
+0.166	7716_6166 9716_33838 7716_6166
+0.140	7755_48439
+0.138	9716_33838 7716_6166 7716_6166
+0.133	7716_6166 7716_6166 9716_33838
… 331 more positive …
… 49660 more negative …

Weight^?	Feature
+0.343	9934_39427
+0.325	9716_33107
+0.243	7752_517
+0.195	9934_39427 7752_517
+0.181	9785_17932
+0.180	9934_45313
+0.165	9785_17921
+0.163	9934_29472
+0.152	9934_51714
+0.148	9716_33102 9716_33107
… 859 more positive …
… 49132 more negative …

Weight^?	Feature
+0.428	7752_766
+0.415	9752_15544
+0.292	7757_2990
+0.275	7752_985
+0.266	9752_11327
+0.206	7752_45684
+0.188	9702_30571
+0.163	7757_2991
+0.136	7752_10367
+0.136	9752_772
… 340 more positive …
… 49651 more negative …

Weight^?	Feature
+0.373	9734_39958
+0.337	9734_48386
+0.327	9734_48406
+0.321	9734_39938
+0.297	9734_48388
+0.164	9935_39681
+0.145	9734_48406 9734_48386
+0.142	9935_14336
+0.141	7752_22712
+0.140	9734_39958 9734_39938
… 820 more positive …
… 49171 more negative …

Weight^?	Feature
+0.462	9746_38506
+0.318	9953_6545
+0.297	9757_13626
+0.225	9745_38058
+0.193	9953_6545 9746_38506
+0.178	9746_38506 9953_6545
+0.162	9953_6544
+0.155	9746_38506 9746_38506
+0.150	9757_13626 9745_38058
+0.147	9757_14429
… 149 more positive …
… 49842 more negative …

Maybe increasing max_features for TF-IDF vectorizer will improve the model.
Also, we see that unigrams and bigrams are used extensively, while trigrams are not that important.

Tuning¶

TF-IDF¶

In [82]:

X_train_tfidf, X_test_tfidf = tf_idf_transform(
    ngram_range=(1, 3),
    max_features=100_000,
    train_texts=train_journeys,
    test_texts=test_journeys,
)

In [83]:

res = fit_and_score(
    C=15,
    X_train=X_train_tfidf,
    y_train=y_train,
    X_val=X_test_tfidf,
    y_val=y_test,
)

print(f"Precision: {res[0]} \nRecall: {res[1]}")

Precision: 0.856 
Recall: 0.898

`ngram_range`¶

In [84]:

ngram_ranges = [(1, 1), (1, 2), (1, 3), (1, 4), (1, 5)]

In [85]:

scores = {}
for i, ngram_range in tqdm(enumerate(ngram_ranges), desc="Tuning ngram ranges"):
    X_train_tfidf, X_test_tfidf = tf_idf_transform(
        ngram_range=ngram_range,
        max_features=100_000,
        train_texts=train_journeys,
        test_texts=test_journeys,
    )
    scores[ngram_range] = fit_and_score(
        C=15,
        X_train=X_train_tfidf,
        y_train=y_train,
        X_val=X_test_tfidf,
        y_val=y_test,
    )
    print(f"({i}) {ngram_range}: {scores[ngram_range]}")

Tuning ngram ranges: 0it [00:00, ?it/s]

(0) (1, 1): (0.837, 0.884)
(1) (1, 2): (0.849, 0.892)
(2) (1, 3): (0.856, 0.898)
(3) (1, 4): (0.856, 0.898)
(4) (1, 5): (0.856, 0.898)

Seems like $(1, 3)$ ngram range is optimal for this task

`max_features`¶

In [86]:

max_features_ = [10_000, 50_000, 100_000, 250_000, 500_000]

In [87]:

scores = {}
for i, max_features in tqdm(
    enumerate(max_features_), desc="Tuning maximum number of features"
):
    X_train_tfidf, X_test_tfidf = tf_idf_transform(
        ngram_range=(1, 3),
        max_features=max_features,
        train_texts=train_journeys,
        test_texts=test_journeys,
    )
    scores[max_features] = fit_and_score(
        C=15,
        X_train=X_train_tfidf,
        y_train=y_train,
        X_val=X_test_tfidf,
        y_val=y_test,
    )
    print(f"({i}) {max_features}: {scores[max_features]}")

Tuning maximum number of features: 0it [00:00, ?it/s]

(0) 10000: (0.83, 0.88)
(1) 50000: (0.85, 0.894)
(2) 100000: (0.856, 0.898)
(3) 250000: (0.853, 0.894)
(4) 500000: (0.853, 0.894)

$\text{100 000}$ looks like an optimal choice

Final params¶

{ngram_range=(1, 3), max_features=100_000}

Logistic Regression¶

Let's split test dataset into validation and testing datasets

In [88]:

X_train_tfidf, X_test_tfidf = tf_idf_transform(
    ngram_range=(1, 3),
    max_features=100_000,
    train_texts=train_journeys,
    test_texts=test_journeys,
)

In [89]:

X_test_tfidf_test, X_test_tfidf_val, y_test_test, y_test_val = train_test_split(
    X_test_tfidf, y_test, test_size=0.5
)

In [90]:

Cs = np.logspace(0, 2, 10)
Cs

Out[90]:

array([  1.        ,   1.66810054,   2.7825594 ,   4.64158883,
         7.74263683,  12.91549665,  21.5443469 ,  35.93813664,
        59.94842503, 100.        ])

In [91]:

scores = {}

for i, C in tqdm(enumerate(Cs), desc="Tuning Logistic Regression"):
    scores[C] = fit_and_score(C, X_train_tfidf, y_train, X_test_tfidf_val, y_test_val)
    print(f"({i}) {C}: {scores[C]}")

Tuning Logistic Regression: 0it [00:00, ?it/s]

(0) 1.0: (0.816, 0.842)
(1) 1.6681005372000588: (0.829, 0.853)
(2) 2.7825594022071245: (0.839, 0.863)
(3) 4.641588833612778: (0.846, 0.871)
(4) 7.742636826811269: (0.846, 0.871)
(5) 12.91549665014884: (0.853, 0.878)
(6) 21.544346900318832: (0.853, 0.878)
(7) 35.93813663804626: (0.853, 0.878)
(8) 59.94842503189409: (0.853, 0.878)
(9) 100.0: (0.853, 0.878)

Seems like, there is no more room for improvement, and we can pick C values from range $(10, 60)$.

Let's test on the part of the test dataset

In [92]:

res = fit_and_score(
    C=15,
    X_train=X_train_tfidf,
    y_train=y_train,
    X_val=X_test_tfidf_test,
    y_val=y_test_test,
)
print(f"Precision: {res[0]} \nRecall: {res[1]}")

Precision: 0.825 
Recall: 0.845

And on the whole test dataset

In [93]:

res = fit_and_score(
    C=15, X_train=X_train_tfidf, y_train=y_train, X_val=X_test_tfidf, y_val=y_test
)
print(f"Precision: {res[0]} \nRecall: {res[1]}")

Precision: 0.856 
Recall: 0.898

The result on the whole test dataset is better because we tuned the algorithm on one of its parts

Final params¶

C=15

Training and Prediction on unlabeled data¶

In [94]:

X_train = test_data_left.drop_duplicates(["hash_id", "user_journey"])
X_test = test_data_right.drop_duplicates(["hash_id", "user_journey"])

y_train = X_train["hash_id"].values
y_test = X_test["hash_id"].values
train_journeys = X_train["user_journey"].values
test_journeys = X_test["user_journey"].values

In [95]:

X_train_tfidf, X_test_tfidf = tf_idf_transform(
    ngram_range=(1, 3),
    max_features=100_000,
    train_texts=train_journeys,
    test_texts=test_journeys,
)

In [96]:

clf = LogisticRegression(
    C=15,
    random_state=SEED,
    n_jobs=-1,
    verbose=2,
    multi_class="multinomial",  # to solve the task like Softmax Classifier
    solver="sag",  # for faster training
)

In [97]:

clf.fit(X_train_tfidf, y_train)

[Parallel(n_jobs=-1)]: Using backend ThreadingBackend with 16 concurrent workers.

convergence after 54 epochs took 384 seconds

[Parallel(n_jobs=-1)]: Done   1 out of   1 | elapsed:  6.4min finished

Out[97]:

LogisticRegression(C=15, multi_class='multinomial', n_jobs=-1, random_state=42,
                   solver='sag', verbose=2)

In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.

In [98]:

y_pred = clf.predict(X_test_tfidf)

Saving results¶

In [99]:

res = pd.DataFrame.from_dict({"id1": y_test, "id2": y_pred})

In [100]:

res.to_csv(FINAL_MAPPING_PATH_CSV, index=False)

Task¶

Imports¶

Paths¶

Input¶

Output¶

Loading Data¶

Dataset Overview¶

Dataset Close Look¶

Reference¶

Data¶

lac¶

cid¶

Fixing CID¶

lac and cid interaction¶

ts¶

fulldate¶

hash_id¶

Data transformation¶

Saving a bit of memory¶

ts and fulldate¶

Transforming ts and fulldate into datetime format¶

Date and Time features extraction¶

Checking data integrity¶

User journeys creation from lac_cid feature¶

Datetime based features¶

Splitting data¶

Splitting¶

Training dataset¶

New ID for training dataset¶

Testing dataset¶

Modelling¶

Preparing data¶

TF-IDF for user journeys¶

Logistic Regression¶

Results analysis¶

Tuning¶

TF-IDF¶

ngram_range¶

max_features¶

Final params¶

Logistic Regression¶

Final params¶

Training and Prediction on unlabeled data¶

Saving results¶

`lac`¶

`cid`¶

`lac` and `cid` interaction¶

`ts`¶

`fulldate`¶

`hash_id`¶

`ts` and `fulldate`¶

Transforming `ts` and `fulldate` into `datetime` format¶

User journeys creation from `lac_cid` feature¶

`ngram_range`¶

`max_features`¶