Low-Rank Boolean Matrix Approximation by Integer Programming
Machine Learning
2018-03-14 v1 Discrete Mathematics
Machine Learning
Abstract
Low-rank approximations of data matrices are an important dimensionality reduction tool in machine learning and regression analysis. We consider the case of categorical variables, where it can be formulated as the problem of finding low-rank approximations to Boolean matrices. In this paper we give what is to the best of our knowledge the first integer programming formulation that relies on only polynomially many variables and constraints, we discuss how to solve it computationally and report numerical tests on synthetic and real-world data.
Cite
@article{arxiv.1803.04825,
title = {Low-Rank Boolean Matrix Approximation by Integer Programming},
author = {Reka Kovacs and Oktay Gunluk and Raphael Hauser},
journal= {arXiv preprint arXiv:1803.04825},
year = {2018}
}