compare_models.py

from sklearn import datasets
from sklearn.model_selection import train_test_split
from sklearn.metrics import mean_squared_error as mse
from sklearn.ensemble import RandomForestRegressor
from lightgbm import LGBMRegressor

X, y = datasets.load_diabetes(return_X_y=True)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)


rf_regressor = RandomForestRegressor(n_estimators=30)
rf_regressor.fit(X_train, y_train)
train_error = mse(y_train, rf_regressor.predict(X_train))
test_error = mse(y_test, rf_regressor.predict(X_test))
print('RF test error:', test_error)
print('RF overfit ratio:', test_error/train_error)

boosting_regressor = LGBMRegressor(n_estimators=30, learning_rate=0.12)
boosting_regressor.fit(X_train, y_train)
train_error = mse(y_train, boosting_regressor.predict(X_train))
test_error = mse(y_test, boosting_regressor.predict(X_test))
print('Boosting test error:', test_error)
print('Boosting overfit ratio:', test_error/train_error)

dart_regressor = LGBMRegressor(n_estimators=30, learning_rate=0.12, boosting_type='dart', skip_drop=0.7)
dart_regressor.fit(X_train, y_train)
train_error = mse(y_train, dart_regressor.predict(X_train))
test_error = mse(y_test, dart_regressor.predict(X_test))
print('Dart test error:', test_error)
print('Dart overfit ratio:', test_error/train_error)

comparison.csv