Tabular example¶

In [ ]:

from fastai import *          # Quick accesss to most common functionality
from fastai.tabular import *  # Quick accesss to tabular functionality     # Access to example data provided with fastai

Tabular data should be in a Pandas DataFrame.

In [ ]:

path = untar_data(URLs.ADULT_SAMPLE)
df = pd.read_csv(path/'adult.csv')

In [ ]:

dep_var = '>=50k'
cat_names = ['workclass', 'education', 'marital-status', 'occupation', 'relationship', 'race']
cont_names = ['age', 'fnlwgt', 'education-num']
procs = [FillMissing, Categorify, Normalize]

In [ ]:

test = TabularList.from_df(df.iloc[800:1000].copy(), path=path, cat_names=cat_names, cont_names=cont_names)

In [ ]:

data = (TabularList.from_df(df, path=path, cat_names=cat_names, cont_names=cont_names, procs=procs)
                           .split_by_idx(list(range(800,1000)))
                           .label_from_df(cols=dep_var)
                           .add_test(test, label=0)
                           .databunch())

In [ ]:

data.show_batch(rows=10)

workclass	education	marital-status	occupation	relationship	race	education-num_na	age	fnlwgt	education-num
Private	HS-grad	Married-civ-spouse	Craft-repair	Husband	Asian-Pac-Islander	False	0.1769	-0.1070	-0.4224
Local-gov	Some-college	Never-married	Prof-specialty	Own-child	White	False	-0.0430	0.4453	-0.0312
Private	Some-college	Married-civ-spouse	Transport-moving	Husband	White	False	0.9098	1.1638	-0.0312
Private	Some-college	Married-civ-spouse	Adm-clerical	Wife	White	False	-0.9959	-1.0879	-0.0312
?	HS-grad	Widowed	?	Unmarried	White	False	2.0093	-0.9140	-0.4224
Private	HS-grad	Divorced	Adm-clerical	Not-in-family	White	False	0.9831	-0.6828	-0.4224
Private	HS-grad	Married-civ-spouse	Craft-repair	Husband	White	False	-0.4828	-1.3949	-0.4224
Private	Bachelors	Never-married	Prof-specialty	Not-in-family	White	False	0.4701	-0.2632	1.1422
Private	Assoc-acdm	Married-civ-spouse	Adm-clerical	Husband	Asian-Pac-Islander	False	0.5434	-0.1371	0.7511
Private	Bachelors	Married-civ-spouse	Exec-managerial	Husband	White	False	-0.0430	0.0111	1.1422

In [ ]:

learn = tabular_learner(data, layers=[200,100], metrics=accuracy)
learn.fit(1, 1e-2)

Total time: 00:04
epoch  train_loss  valid_loss  accuracy
1      0.373440    0.381403    0.825000  (00:04)

Inference¶

In [ ]:

row = df.iloc[0]

In [ ]:

learn.predict(row)

Out[ ]:

(1, tensor(0), tensor([0.5647, 0.4353]))

In [ ]: