Logistic Regression CV

lr.py

from sklearn.ensemble import RandomForestClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.cross_validation import train_test_split
from sklearn.metrics import accuracy_score, roc_auc_score


def score(model, X, y):
    X_train, X_test, y_train, y_test = train_test_split(X, y)
    model.fit(X_train, y_train)
    pred_proba = model.predict_proba(X_test)
    pred = model.predict(X_test)
    from sklearn.metrics import accuracy_score, roc_auc_score
    print('acc:', accuracy_score(y_test, pred))
    print('auc:', roc_auc_score(y_test, pred_proba[:, 1]))
    return model, roc_auc_score(y_test, pred_proba[:, 1])
  
lr = LogisticRegression(class_weight='balanced')
score(lr, X, y)

rfc = RandomForestClassifier(class_weight='balanced')
model = score(rfc, X, y)