gbm_employee.py

# Build Gradient Boosting classifier
pip_gb = make_pipeline(StandardScaler(),
                       GradientBoostingClassifier(loss="deviance",
                                                  random_state=123))
hyperparam_grid = {"gradientboostingclassifier__max_features": ["log2", 0.5],
                   "gradientboostingclassifier__n_estimators": [100, 300, 500],
                   "gradientboostingclassifier__learning_rate": [0.001, 0.01, 0.1],
                   "gradientboostingclassifier__max_depth": [1, 2, 3]}
gs_gb = GridSearchCV(pip_gb,
                      param_grid=hyperparam_grid,
                      scoring="f1",
                      cv=10,
                      n_jobs=-1)
gs_gb.fit(X_train, y_train)
print(f"\033[1m\033[0mThe best hyperparameters:\n{'-' * 25}")
for hyperparam in gs_gb.best_params_.keys():
    print(hyperparam[hyperparam.find("__") + 2:], ": ", gs_gb.best_params_[hyperparam])
print(f"\033[1m\033[94mBest 10-folds CV f1-score: {gs_gb.best_score_ * 100:.2f}%.")