Skip to main content

KNeighborsRegressor

#include <Skigen/Neighbors>

template <typename Scalar = double>
class Skigen::KNeighborsRegressor(n_neighbors=5)

Regression based on k-nearest neighbors.

The target is predicted by local interpolation of the targets associated of the nearest neighbors in the training set (mean of the k nearest values).

Mirrors sklearn.neighbors.KNeighborsRegressor.

Parameters:

  • n_neighbors : int, default=5 Number of neighbors to use (int, default 5).

Attributes:

  • n_neighbors : int Number of neighbors.

  • n_targets : int

Methods​

fit(X, y)​

Fit the k-nearest neighbors regressor from the training set.

predict(X)​

Predict target values for the provided data.

score(X, y)​

Return the R2R^2 coefficient of determination.

fit_multi(X, Y)​

Fit on multi-target Y (shape n × n_targets).

Stores X and Y; prediction averages neighbour rows of Y_train_ per target. Single-target API is unchanged — predict() keeps returning a 1-D vector reflecting the first target column when fit_multi was used.

predict_multi(X)​

Predict (n_samples × n_targets) — mean of neighbour rows from Y_train_ per target.

Example​

// KNN for regression
Eigen::VectorXd y_reg(split.X_train.rows());
for (Eigen::Index i = 0; i < y_reg.size(); ++i)
    y_reg(i) = split.X_train(i, 0) + 0.5 * split.X_train(i, 1);

Skigen::KNeighborsRegressor<double> knn_reg(5);
knn_reg.fit(split.X_train, y_reg);

Eigen::VectorXd y_reg_test(split.X_test.rows());
for (Eigen::Index i = 0; i < y_reg_test.size(); ++i)
    y_reg_test(i) = split.X_test(i, 0) + 0.5 * split.X_test(i, 1);

std::cout << "\n=== KNeighborsRegressor (k=5) ===\n";
std::cout << "R²: " << knn_reg.score(split.X_test, y_reg_test) << "\n";