Anaconda Defect

logistic_regression.py

import numpy as np
import pandas as pd

from nose.tools import *

from sklearn.linear_model import LogisticRegression
from sklearn.preprocessing import PolynomialFeatures

bank_data = pd.read_csv('./data/bank.csv', sep=";")
bank_features = bank_data.drop('y', axis=1)
bank_features = pd.get_dummies(bank_features)
bank_output = bank_data.loc[:,'y']

quad_feature_transformer = PolynomialFeatures(degree=2, interaction_only=True)
bank_features_quad = quad_feature_transformer.fit_transform(bank_features)

bank_model_quad = LogisticRegression(C=1e6)
bank_model_quad.fit(bank_features_quad, bank_output)
accuracy_score_quad = bank_model_quad.score(bank_features_quad, bank_output)

print("Accuracy: {}".format(accuracy_score_quad)) 
# Ubuntu output:  0.90400353904
# Windows output: 0.89869497895

assert_almost_equal(accuracy_score_quad, 0.8986949789869498, delta = 0.001)