Skip to content
Quasi-Monte Carlo Feature Maps for Gaussian Kernel
MATLAB
Branch: master
Clone or download

Latest commit

Fetching latest commit…
Cannot retrieve the latest commit at this time.

Files

Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
LICENSE
README.md
demo.m
qmc_feature_maps_complex.m

README.md

QMC Feature Maps

Codes for quasi-Monte Carlo feature maps for Gaussian kernels in MATLAB.

About

Given any pair of points x and y, compute a feature map z(x) and z(y) such that k(x,y) is approximated by the inner product between z(x) and z(y) where in this case k is the Gaussian kernel.

Usage

See qmc_faeture_maps_complex.m for usage. See demo.m for a simple example.

References

J. Yang, V. Sindhwani, H. Avron, and M. W. Mahoney, Quasi-Monte Carlo Feature Maps for Shift-Invariant Kernels. International Conference on Machine Learning (ICML), 2014.

You can’t perform that action at this time.