This is a Java library that provides some algorithms to efficiently perform principal component analysis.
The library can be downloaded as a binary or source package from https://github.com/trickl/trickl-pca/downloads.
The source can be also retrieved straight from the Git repository: git clone git@github.com:trickl/trickl-pca.git
.
HallMarshallMartin - An incremental algorithm described by a paper by Hall, Marshall and Martin. Modified beyond the scope of that paper to work efficienttly on sparse matrices. SVDPCA - A standard batch algorithm for PCA that requires a Singular Value Decomposition solver. SkocajLeonardis - An incremental algorithm that allows each input to have a temporal weight, based on a paper by Skocaj and Leonardis.
Depends on the Colt matrix library.
The program is available under GNU General Public License (see file license-gpl.txt
in the distribution package for the text of the license).
Some Junit tests are provided and should be used to consult usage.