Overview
The RidgeRegression class implements ridge regression with L2 regularization. It uses standard C++ vectors for data representation.
Template parameter:
T - Numeric type (typically float or double). Defaults to double.
Constructor
L2 regularization parameter λ. Higher values increase regularization strength
Methods
fit
void fit(const std::vector<std::vector<T>>& X,
const std::vector<T>& y)
X
std::vector<std::vector<T>>
Feature matrix where each inner vector represents one sample’s features
Target vector with one value per sample
predict
std::vector<T> predict(const std::vector<std::vector<T>>& X) const
X
std::vector<std::vector<T>>
Feature matrix for samples to predict
Predicted values, one per sample
weights
const std::vector<T>& weights() const
set_lambda
void set_lambda(T lambda)
New L2 regularization parameter value
get_lambda
Current L2 regularization parameter value
Usage example
#include <ridge_regression.h>
#include <vector>
#include <iostream>
// Create training data
std::vector<std::vector<double>> X = {
{1.0, 2.0, 3.0},
{2.0, 3.0, 4.0},
{3.0, 4.0, 5.0},
{4.0, 5.0, 6.0}
};
std::vector<double> y = {2.5, 4.5, 6.5, 8.5};
// Create model with lambda = 0.5
RidgeRegression<double> model(0.5);
// Fit the model
model.fit(X, y);
// Make predictions
std::vector<std::vector<double>> X_test = {
{5.0, 6.0, 7.0},
{6.0, 7.0, 8.0}
};
std::vector<double> predictions = model.predict(X_test);
for (size_t i = 0; i < predictions.size(); ++i) {
std::cout << "Prediction " << i << ": " << predictions[i] << std::endl;
}
// Access weights
const auto& w = model.weights();
// Update lambda
model.set_lambda(1.0);
model.fit(X, y); // Refit with new lambda
min_w ||Xw - y||² + λ||w||²