max_features=None)
# Fit classifier to training set
clf = clf.fit(X_train, y_train, sample_weight=sample_weight)
# In[ ]:
# Compute training metrics
accuracy = clf.score(X_train, y_train)
# Predict labels of test set
train_pred = clf.predict(X_train)
# Compute MSE, confusion matrix, classification report
mse = mean_squared_error(y_train, train_pred)
conf_mat = confusion_matrix(y_train.round(), train_pred.round())
clas_rep = classification_report(y_train.round(), train_pred.round())
# Print reports
print('{:=^80}'.format('RF training report'))
print('Accuracy: %.4f' % accuracy)
print("MSE: %.4f" % mse)
print("Confusion matrix:\n{}".format(conf_mat))
print("Classification report:\n{}".format(clas_rep))
# Here is how to read the above confusion matrix:
#
# | | Prediction: 0 (Unaffected) | Prediction: 1 (BA) | Prediction: 2 (PCL) |
# |-----------|--------------------------------------|----------------------------|------------------------------|
# | Actual: 0 (Unaffected) | __Unaffected classified as unaffected__ | Unaffected classified as BA | Unaffected classified as PCL |
# | Actual: 1 (BA) | BA classified as unaffected | __BA classified as BA__ | BA classified as PCL |
