LightGBM

param = {
    'learning_rate' : 0.1,
    'max_depth' : 5,
    'min_child_samples' : 10,
    'n_estimators' : 20,
    'num_leaves' : 5,
    'reg_alpha' : 0.1,
}
lgb_base = lgb.LGBMRegressor(**param).fit(X_train, y_train)

def target(
    learning_rate,
    max_depth,
    min_child_samples,
    n_estimators,
    num_leaves,
    reg_alpha,
):
    param = {
    'learning_rate' : learning_rate,
    'max_depth' : int(max_depth),
    'min_child_samples' : int(min_child_samples),
    'n_estimators' : int(n_estimators),
    'num_leaves' : int(num_leaves),
    'reg_alpha' : reg_alpha,
    }
    model = lgb.LGBMRegressor(**param).fit(X_train, y_train)
    y_val_pred = model.predict(X_val)
    return -mean_squared_error(y_val, y_val_pred)

search_range={
    'learning_rate' : (0.01, 0.3),
    'max_depth' : (1,10),
    'min_child_samples' : (5,50),
    'n_estimators' : (1, 50),
    'num_leaves' : (3, 50),
    'reg_alpha' : (0.01, 0.6),
}
search_engine = BayesianOptimization(target, search_range, verbose=1)
search_engine.maximize(
    init_points = 50,
    n_iter = 300,
)

params_bo = search_engine.max['params']
params_bo['max_depth'] = int(round(params_bo['max_depth']))
params_bo['num_leaves'] = int(round(params_bo['num_leaves']))
params_bo['n_estimators'] = int(round(params_bo['n_estimators']))
params_bo['min_child_samples'] = int(round(params_bo['min_child_samples']))
params_bo

lgb_bo = lgb.LGBMRegressor(**params_bo).fit(X_in, y_in)