Module multi_scale_rbf 

Multi-scale RBF kernel approximation methods

This module implements multi-scale RBF kernel approximation that uses multiple bandwidth parameters to capture patterns at different scales. This provides better approximation quality for data with features at multiple scales.

Structs

MultiScaleRBFSampler

Multi-scale RBF sampler that generates random Fourier features at multiple scales

Enums

BandwidthStrategy

Multi-scale bandwidth selection strategy BandwidthStrategy

CombinationStrategy

Feature combination strategy for multi-scale features CombinationStrategy