Module structured_random_features 

Structured orthogonal random features for efficient kernel approximation

This module implements structured random feature methods that use structured random matrices (like Hadamard matrices) to reduce computational complexity while maintaining approximation quality. Also includes quasi-random and low-discrepancy sequence methods for improved feature distribution.

Structs

FastWalshHadamardTransform

Fast Walsh-Hadamard Transform for efficient structured features

LowDiscrepancySequences

Low-discrepancy sequence generators for quasi-random features

StructuredRFFHadamard

Structured Random Features using Fast Walsh-Hadamard Transform

StructuredRandomFeatures

Structured orthogonal random features

Enums

QuasiRandomSequence

Type of quasi-random sequence to use for feature generation QuasiRandomSequence

StructuredMatrix

Type of structured matrix to use StructuredMatrix