init_params = {
"boosting_type": "gbdt",
"objective": "regression",
"metric": "mae",
"n_estimators": 500,
"learning_rate": 0.003,
"num_leaves": 128,
"feature_fraction": 0.4,
# "bagging_fraction": 1,
# "bagging_freq": 0,
"min_data_in_leaf": 500,
"min_sum_hessian_in_leaf": 1,
# "lambda_l1": 0,
# "lambda_l2": 0,
"importance_type": "gain",
"n_jobs": 4,
"random_state": 456,
}
fit_params = {"callbacks": [counter()]}
model_run(lgb.LGBMRegressor,
2017,
init_params,
fit_params=fit_params,
pca=True,
save_model=True,
out="results/lgbm_3")
print("Done")
"learning_rate": 0.002,
"num_leaves": 128,
"feature_fraction": 0.32,
"bagging_fraction": 0.5,
"bagging_freq": 1,
"min_data_in_leaf": 500,
"min_sum_hessian_in_leaf": 1,
# "lambda_l1": 0,
# "lambda_l2": 0,
"importance_type": "gain",
"n_jobs": -1,
"random_state": 456,
}
fit_params = {"callbacks": [counter()]}
model_run(lgb.LGBMRegressor,
2016,
init_params,
fit_params=fit_params,
save_model=True,
out="results/lrf_2")
model_run(lgb.LGBMRegressor,
2017,
init_params,
fit_params=fit_params,
save_model=True,
out="results/lrf_2")
print("Done")