from sklearn.linear_model import LinearRegression, HuberRegressor, RidgeCV
from sklearn.ensemble import RandomForestRegressor
from sklearn.model_selection import GridSearchCV, LeaveOneOut

alphas_cv = [0.05, 0.1, 0.5, 1.0, 5.0, 10.0, 50.0, 100.0]

rf = RandomForestRegressor(random_state=42)
loocv = LeaveOneOut()

# Define strict parameters for RandomForest to prevent overfitting
rf_params = {
    'n_estimators': [50, 100], 
    'max_depth': [2, 3, 4],
    'min_samples_leaf': [3, 5, 7], # Crucial for small datasets
    'min_samples_split': [4, 6, 8],
    'max_features': ['sqrt', 'log2', 0.5], # Increases tree diversity
}

models = {
    "lr": LinearRegression(),
    "huber": HuberRegressor(max_iter=5000, tol=1e-3, alpha=0.001),
    "ridgecv": RidgeCV(alphas=alphas_cv),
    "rf": GridSearchCV(
        estimator=rf,
        param_grid=rf_params,
        cv=loocv,
        scoring='neg_mean_squared_error',
        verbose=1,
        n_jobs=-1
    ),
}