An algorithm to (approximately) solve sparse and nonnegative PCA quadratic maximization problems
The algorithm operates on the raw data (samples x features matrix) or the covariance matrix.
It is based on Algorithm 3 in Nonnegative Sparse PCA with Provable Guarantees (ICML 2014).
See QuickDemo.m for an example on how to use the code.
The log4m.m created by Luke Winslow and is available through Matlab Central, here). The file is distributed under the BSD license.