/home/runner/work/deep-learning-algo-impls/deep-learning-algo-impls/include/ml/pca.hpp

A class implementing Principal Component Analysis (PCA)

A class implementing Principal Component Analysis (PCA)This class provides functionality for dimensionality reduction using PCA. It computes principal components of the input data and allows projection onto a lower-dimensional space.

Template Parameters

T	The data type for matrix elements (typically float or double)

// Create a PCA object
PCA<double> pca;
 
// Fit PCA to data
Matrix<double> data = your_data_matrix;
pca.fit(data);
 
// Transform data to reduced dimensions
Matrix<double> transformed = pca.transform(data, 2); // Reduce to 2 dimensions

#pragma once
 
#include <algorithm>
#include <cmath>
#include <stdexcept>
#include <vector>
#include "../utils/matrix.hpp"
 
namespace ml {
    using namespace utils;
    template<typename T>
    class PCA {
    public:
        PCA() = default;
 
        void fit(const Matrix<T> &data, bool center = true, bool scale = false);
 
        Matrix<T> transform(const Matrix<T> &data, size_t n_components = 0) const;
 
        Matrix<T> fit_transform(const Matrix<T> &data, size_t n_components = 0, bool center = true, bool scale = false);
 
        std::vector<T> explained_variance_ratio() const;
 
        Matrix<T> components() const;
 
        std::vector<T> singular_values() const;
 
    private:
        Matrix<T> components_; 
        std::vector<T> singular_values_; 
        std::vector<T> explained_variance_; 
        std::vector<T> explained_variance_ratio_; 
        std::vector<T> mean_; 
        std::vector<T> scale_; 
        bool is_fitted_ = false; 
    };
 
    // Type aliases for common use cases
    using PCAF = PCA<float>;
    using PCAD = PCA<double>;
 
} // namespace ml