The program takes a binary matrix as an input and factorizes it into two binary matrices such that their Boolean matrix product is an approximation of the input matrix.
The optimization uses an annealed EM scheme to maximize the likelihood of the latent feature model described in our paper.
Originally, we developed our model and code to solve the role mining problem from the field of information theory. However, it could be used to approximately factorize any binary matrix. This is useful for database tiling, the set basis problem, market basket analysis and other problems.
Please contact us if you have comments or questions.
- Changes to previous version:
Initial Announcement on mloss.org.
Leave a comment
You must be logged in to post comments.