churn_result = churn_df['link']
churn_df.drop('link',axis=1,inplace=True)
churn_train = churn_df.tail(-500)
churn_train_result = churn_result.tail(-500)
churn_test = churn_df.head(500)
# Save test to a csv file
churn_test.to_csv("test_churn.csv")
dv = DictVectorizer()
dv.fit(churn_df.T.to_dict().values())
from sklearn.preprocessing import StandardScaler
yes_no_cols = ["price"]
scaler = StandardScaler()
# Feature matrix
X = dv.as_matrix().astype(np.float)
X = scaler.fit_transform(X)
# Target values
y = np.where(churn_result == 'True.',1,0)
