Scikit-Learn GridSearchCV MLPRegressor

Hyperparameter tuning is essential for optimizing machine learning models for the best performance. In this example, we’ll demonstrate how to use scikit-learn’s GridSearchCV to perform hyperparameter tuning for MLPRegressor, a popular algorithm for regression tasks.

Grid search is a method for evaluating different combinations of model hyperparameters to find the best performing configuration. It exhaustively searches through a specified parameter grid, trains and evaluates the model for each combination using cross-validation, and selects the hyperparameters that yield the best performance metric.

MLPRegressor is a multi-layer perceptron regressor used for regression tasks. It relies on a feedforward artificial neural network model to predict continuous values based on input features. The model is trained by minimizing the mean squared error loss function.

The key hyperparameters for MLPRegressor include the hidden_layer_sizes, which defines the architecture of the neural network (number of layers and neurons per layer); the activation, which specifies the activation function for the hidden layers; the solver, which is the optimization algorithm for weight optimization; and alpha, the regularization term to prevent overfitting.

from sklearn.datasets import make_regression
from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.neural_network import MLPRegressor

# Generate synthetic regression dataset
X, y = make_regression(n_samples=1000, n_features=10, noise=0.1, random_state=42)

# Split into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Define parameter grid
param_grid = {
    'hidden_layer_sizes': [(50,), (100,), (50, 50)],
    'activation': ['relu', 'tanh'],
    'solver': ['adam', 'lbfgs'],
    'alpha': [0.0001, 0.001, 0.01]
}

# Perform grid search
grid_search = GridSearchCV(estimator=MLPRegressor(random_state=42, max_iter=100),
                           param_grid=param_grid,
                           cv=5,
                           scoring='neg_mean_squared_error')
grid_search.fit(X_train, y_train)

# Report best score and parameters
print(f"Best score: {grid_search.best_score_:.3f}")
print(f"Best parameters: {grid_search.best_params_}")

# Evaluate on test set
best_model = grid_search.best_estimator_
mse = -best_model.score(X_test, y_test)
print(f"Test set mean squared error: {mse:.3f}")

Running the example gives an output like:

Best score: -0.625
Best parameters: {'activation': 'relu', 'alpha': 0.0001, 'hidden_layer_sizes': (50, 50), 'solver': 'lbfgs'}
Test set mean squared error: -1.000

The steps are as follows:

1. Generate a synthetic regression dataset using scikit-learn’s make_regression function.
2. Split the dataset into train and test sets using train_test_split.
3. Define the parameter grid with different values for hidden_layer_sizes, activation, solver, and alpha hyperparameters.
4. Perform grid search using GridSearchCV, specifying the MLPRegressor model, parameter grid, 5-fold cross-validation, and negative mean squared error scoring metric.
5. Report the best cross-validation score and best set of hyperparameters found by grid search.
6. Evaluate the best model on the hold-out test set and report the mean squared error.

By using GridSearchCV, we can easily explore different hyperparameter settings and find the combination that maximizes the model’s performance. This automated approach saves time and effort compared to manual hyperparameter tuning and helps ensure we select the best configuration for our MLPRegressor model.