On the applicability of the Hadamard as an input modulator for problems of classification

The library explores the applicability of Hadamard as an input modulator for problems classification. It introduces a framework in C++ to use kernel approximates mini-batch setting with Stochastic Gradient Descent. algorithm requires compute product matrices Walsh Hadamard. A free-standing cache friendly Fast that achieves compelling speed is provided, well lightweight efficient CPU implementation method research and practical purposes alike.