Skip to main content

Normalizer

#include <Skigen/Preprocessing>

template <typename Scalar = double>
class Skigen::Normalizer(norm=Norm::L2)

Normalize samples individually to unit norm.

Each sample (i.e. each row of the data matrix) with at least one non-zero component is rescaled independently of other samples so that its norm (L1, L2, or max) equals one.

Mirrors sklearn.preprocessing.Normalizer.

Parameters:

  • norm : Norm, default=Norm::L2 The norm to use (Norm, default Norm::L2).

Attributes:

  • norm : Norm The norm type in use.

Methods​

fit(X)​

Fit the Normalizer (stateless — only records n_features_in_).

Parameters:

  • X : MatrixType Training data of shape (n_samples, n_features).

Returns:

  • result : Normalizer Reference to the fitted transformer (*this).

transform(X)​

Normalize each sample to unit norm.

Parameters:

  • X : MatrixType Data matrix of shape (n_samples, n_features).

Returns:

  • result : MatrixType Normalized data of same shape.

Throws:

  • std::runtime_error — if the model has not been fitted.

inverse_transform()​

Not supported — normalization is not reversible.

Throws:

  • std::runtime_error — Always throws.

transform_inplace(X)​

Normalize samples in-place to unit norm.

Parameters:

  • X : Eigen::Ref< MatrixType > Data matrix of shape (n_samples, n_features), modified in place.

Throws:

  • std::runtime_error — if the model has not been fitted.

Example​

// L2 normalization (default) — each row has unit Euclidean norm
Skigen::Normalizer<double> l2_norm(Skigen::Norm::L2);
Eigen::MatrixXd Z_l2 = l2_norm.fit_transform(X);

std::cout << "L2 normalized:\n" << Z_l2 << "\n";
std::cout << "Row norms: ";
for (Eigen::Index i = 0; i < Z_l2.rows(); ++i)
    std::cout << Z_l2.row(i).norm() << " ";
std::cout << "\n\n";

// L1 normalization — each row sums to 1 (in absolute value)
Skigen::Normalizer<double> l1_norm(Skigen::Norm::L1);
Eigen::MatrixXd Z_l1 = l1_norm.fit_transform(X);

std::cout << "L1 normalized:\n" << Z_l1 << "\n";