Explaining a classfier¶

It's good to able to explain what an algorithm is doing (esp in public sector).

It boosts confidence in the ML, good for transparency, and is useful in iterating the model.

This uses the python package lime (Local Interpretable Model-Agnostic Explanations), an implementation by the authors of their paper - https://arxiv.org/abs/1602.04938

In [1]:

! pip install lime

Requirement already satisfied: lime in c:\users\leungr\appdata\local\continuum\anaconda3\lib\site-packages
Requirement already satisfied: scikit-image>=0.12 in c:\users\leungr\appdata\local\continuum\anaconda3\lib\site-packages (from lime)
Requirement already satisfied: numpy in c:\users\leungr\appdata\local\continuum\anaconda3\lib\site-packages (from lime)
Requirement already satisfied: scikit-learn>=0.18 in c:\users\leungr\appdata\local\continuum\anaconda3\lib\site-packages (from lime)
Requirement already satisfied: scipy in c:\users\leungr\appdata\local\continuum\anaconda3\lib\site-packages (from lime)
Requirement already satisfied: six>=1.7.3 in c:\users\leungr\appdata\local\continuum\anaconda3\lib\site-packages (from scikit-image>=0.12->lime)
Requirement already satisfied: networkx>=1.8 in c:\users\leungr\appdata\local\continuum\anaconda3\lib\site-packages (from scikit-image>=0.12->lime)
Requirement already satisfied: pillow>=2.1.0 in c:\users\leungr\appdata\local\continuum\anaconda3\lib\site-packages (from scikit-image>=0.12->lime)
Requirement already satisfied: dask[array]>=0.5.0 in c:\users\leungr\appdata\local\continuum\anaconda3\lib\site-packages (from scikit-image>=0.12->lime)
Requirement already satisfied: decorator>=3.4.0 in c:\users\leungr\appdata\local\continuum\anaconda3\lib\site-packages (from networkx>=1.8->scikit-image>=0.12->lime)
Requirement already satisfied: olefile in c:\users\leungr\appdata\local\continuum\anaconda3\lib\site-packages (from pillow>=2.1.0->scikit-image>=0.12->lime)
Requirement already satisfied: toolz>=0.7.2 in c:\users\leungr\appdata\local\continuum\anaconda3\lib\site-packages (from dask[array]>=0.5.0->scikit-image>=0.12->lime)

In [2]:

import sklearn
import sklearn.ensemble
import numpy as np
import lime
import lime.lime_tabular
from __future__ import print_function
np.random.seed(2)

In [3]:

blob_account_name = "parlpublic"
blob_account_key = "xKEIV42ZsO8eL2IPjvbLarR2Xu1brxGucDauvVytPXD1uKhAfYUId7SwbGF82FslfkKebPB/ic6/RcPYnNBO6w==" 
container = "trainingdata" 
blobname = "5000_edms_justonetopic.csv"
datafile = "output.txt" 
import os
import pandas as pd
from azure.storage.blob import BlockBlobService

dirname = os.getcwd()

blob_service = BlockBlobService(account_name=blob_account_name,account_key=blob_account_key)
blob_service.get_blob_to_path(container, blobname, datafile)

edm = pd.read_csv(datafile, header = 0)
os.remove(os.path.join(dirname, datafile))
print(edm.shape)
edm.head()

(5000, 2)

Out[3]:

	doc	topic
0	That this House notes the conviction in the Un...	House of Lords
1	That this House congratulates the US corporate...	North America
2	That this House is dismayed by the report in t...	Speaker
3	That this House notes that, according to a No....	Members of the Lords
4	That this House notes the appointment of a new...	House of Lords

In [4]:

# load in tag hierarchy to extract top terms
import json
with open("tag_hierarchy.json", 'r') as f:
    data = json.load(f) 

In [5]:

# Cleaning
import re

def clean_text(string):
    string = re.sub(r"\d", "", string) # remove numbers  
    string = re.sub(r"_+", "", string) # remove consecutive underscores
    string = re.sub(r"<[^>]*>", " ", string) #remove all html tags
    string = re.sub(r"[^0-9a-zA-Z]+", " ", string)  # remove speacial chars
    string = string.lower() # tranform to lower case    
    
    return string.strip()

edm["doc"] = edm.doc.apply(clean_text)

def get_parent(term):
    for i in data['children']:
        for k in i['children']:
            if term == k['name']:
                return i['name']
            if "children" in k:
                for j in k['children']:
                    if term == j['name']:
                        return get_parent(k['name'])
            
# get_parent("Supported housing")

edm["topic"] = edm.topic.apply(get_parent)

edm.head()

Out[5]:

	doc	topic
0	that this house notes the conviction in the un...	Parliament, government and politics
1	that this house congratulates the us corporate...	International affairs
2	that this house is dismayed by the report in t...	Parliament, government and politics
3	that this house notes that according to a no d...	Parliament, government and politics
4	that this house notes the appointment of a new...	Parliament, government and politics

In [6]:

from sklearn.feature_extraction.text import CountVectorizer
tf_vectorizer = CountVectorizer(max_df=0.95, min_df=2,max_features=1000,stop_words='english')
tf_vectorizer

Out[6]:

CountVectorizer(analyzer='word', binary=False, decode_error='strict',
        dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',
        lowercase=True, max_df=0.95, max_features=1000, min_df=2,
        ngram_range=(1, 1), preprocessor=None, stop_words='english',
        strip_accents=None, token_pattern='(?u)\\b\\w\\w+\\b',
        tokenizer=None, vocabulary=None)

In [7]:

X = tf_vectorizer.fit_transform(edm['doc'])
X

Out[7]:

<5000x1000 sparse matrix of type '<class 'numpy.int64'>'
	with 150811 stored elements in Compressed Sparse Row format>

In [8]:

doc_array = X.toarray()  
doc_array.astype(float)

Out[8]:

array([[ 0.,  0.,  0., ...,  0.,  0.,  0.],
       [ 0.,  0.,  0., ...,  0.,  0.,  0.],
       [ 0.,  0.,  0., ...,  0.,  0.,  0.],
       ..., 
       [ 0.,  0.,  0., ...,  0.,  1.,  0.],
       [ 0.,  0.,  0., ...,  0.,  0.,  0.],
       [ 0.,  0.,  0., ...,  0.,  0.,  0.]])

In [9]:

analyze = tf_vectorizer.build_analyzer()
analyze("we shall fight on the beaches")

Out[9]:

['shall', 'fight', 'beaches']

In [10]:

feature_names = tf_vectorizer.get_feature_names()

In [11]:

train, test, labels_train, labels_test = sklearn.model_selection.train_test_split(doc_array, edm.topic, train_size=0.80)
rf = sklearn.ensemble.RandomForestClassifier(n_estimators=500)
rf.fit(train, labels_train)

Out[11]:

RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',
            max_depth=None, max_features='auto', max_leaf_nodes=None,
            min_impurity_split=1e-07, min_samples_leaf=1,
            min_samples_split=2, min_weight_fraction_leaf=0.0,
            n_estimators=500, n_jobs=1, oob_score=False, random_state=None,
            verbose=0, warm_start=False)

In [12]:

sklearn.metrics.accuracy_score(labels_test, rf.predict(test))

Out[12]:

0.57399999999999995

In [13]:

from sklearn import preprocessing
le = preprocessing.LabelEncoder()
le.fit(edm.topic)
classes_names = le.classes_

In [14]:

explainer = lime.lime_tabular.LimeTabularExplainer(train, feature_names=feature_names, class_names=classes_names, discretize_continuous=True)

C:\Users\leungr\AppData\Local\Continuum\Anaconda3\lib\site-packages\sklearn\utils\validation.py:429: DataConversionWarning: Data with input dtype int64 was converted to float64 by StandardScaler.
  warnings.warn(msg, _DataConversionWarning)

In [15]:

i = np.random.randint(0, len(test))
tf_vectorizer.inverse_transform(test[i])

Out[15]:

[array(['calls', 'chief', 'concern', 'court', 'engage', 'ensure',
        'following', 'home', 'legislation', 'notes', 'officers', 'order',
        'police', 'properly', 'public', 'recent', 'secretary', 'senior',
        'settlement'], 
       dtype='<U15')]

In [16]:

exp = explainer.explain_instance(test[i], rf.predict_proba, num_features=10, top_labels=3)
exp.show_in_notebook(show_table=True, show_all=False)