Cristian Rusu,
February, 2019

--------------------------------

This package contains proof-of-concept source code for the results presented in: Fast approximations of orthogonal matrices: efficient PCA projections, C. Rusu and L. Rosasco

This software is released under the MIT license.
Tested on Matlab R2018a.
For problems, bug reports, comments and other contact C. Rusu (cristian.rusu@iit.it).

--------------------------------

Starting points:

testRandomOrthonormalMatrices.m - Tests the approximation algorithm on a random full (d x d) orthonormal basis.
testFastPCA_USPSdataset.m - Tests the approximation algorithm on PCA projections extracted from the USPS dataset.