optuna を用いたハイパーパラメータチューニング

[Inazuma110]

こんな感じでチューニングできます．

def objective(trial):
    params = {
        'objective': 'multi:softmax',
        'num_class': 11,
        'max_depth': trial.suggest_int('max_depth', 1, 20),
        'min_child_weight': trial.suggest_int('min_child_weight', 1, 20),
        'subsample': trial.suggest_discrete_uniform('subsample', 0.5, 0.9, 0.1),
        'colsample_bytree': trial.suggest_discrete_uniform('colsample_bytree', 0.5, 0.9, 0.1),
        'n_estimators': trial.suggest_int('n_estimators', 10, 1000),
        'learning_rate': trial.suggest_loguniform('learning_rate', 1e-8, 1.0),
        'reg_lambda': trial.suggest_discrete_uniform('lambda', 0, 1, 0.1)
    }
    model = xgb.XGBClassifier(**params)
    model.fit(train_train_x, train_train_y)

    pred = model.predict(train_test_x)

    accuracy = f1_score(train_test_y, pred, average='micro')
    return (1-accuracy)


study = optuna.create_study()
study.optimize(objective, n_trials=300)

書き方が古いのかwarningが出ます．