Regression with Histogram Gradient Boosting in Scikit-Learn
Python
A step-by-step example of how to train a Histogram Gradient Boosted model using HistGradientBoostingRegressor from Sklearn.
1| from sklearn.ensemble import HistGradientBoostingRegressor 2| from sklearn.metrics import mean_squared_error, mean_absolute_error, max_error, explained_variance_score, mean_absolute_percentage_error 3| 4| # Step 1: Create list containing the indices of categorical features 5| categorical_features = [1] 6| 7| # Step 2: Initialise histogram gradient boosting regression model and fit 8| model = HistGradientBoostingRegressor( loss='squared_error', 9| learning_rate=0.1, 10| max_depth=None, 11| max_bins=255, 12| categorical_features=categorical_features, 13| random_state=101) 14| model.fit(X_train, y_train) 15| 16| # Step 3: make prediction for test data & evaluate performance 17| y_pred = model.predict(X_test) 18| print('RMSE:',mean_squared_error(y_test, y_pred, squared = False)) 19| print('MAE:',mean_absolute_error(y_test, y_pred)) 20| print('MAPE:',mean_absolute_percentage_error(y_test, y_pred)) 21| print('Max Error:',max_error(y_test, y_pred)) 22| print('Explained Variance Score:',explained_variance_score(y_test, y_pred))
133
121
117
109