def evaluate_results(pred_test, pred_train, labels_test, labels_train, name):
accuracy_test = accuracy_score(labels_test, pred_test)
accuracy_train = accuracy_score(labels_train, pred_train)
fig = prettyPicture(clf, features_test, labels_test)
save_fig(fig, name, 'Decision_Tree/images')
print(
name, 'test accuracy {}:, train accuracy {}:'.format(
round(accuracy_test, 3), round(accuracy_train, 3)))
print('confusion_matrix {}:'.format(
confusion_matrix(labels_test, pred_test)))
clf.fit(features_train, labels_train)
# predict train and test
pred_train = clf.predict(features_train)
pred_test = clf.predict(features_test)
# calculate accuracy
accuracy_test = accuracy_score(labels_test, pred_test)
accuracy_train = accuracy_score(labels_train, pred_train)
print(
name,
'test accuracy {}:, train accuracy {}:'.format(round(accuracy_test, 3),
round(accuracy_train, 3)))
# plot the classes
fig = prettyPicture(clf, features_test, labels_test)
save_fig(fig, name, 'Decision_Tree/images')
# confusion matrix
print('confusion_matrix {}:'.format(confusion_matrix(labels_test, pred_test)))
print(classification_report(labels_test, pred_test, labels=[0, 1]))
# plot roc curve
y_test_pred_prob = clf.predict_proba(features_test)[:, 1]
# IMPORTANT: first argument is true values, second argument is predicted probabilities
fpr, tpr, thresholds = roc_curve(labels_test, pred_test)
plt.figure()
plt.plot(fpr, tpr)
plt.xlim([0.0, 1.0])