An Optimization Approach for Solving Orthogonal Nonnegative Matrix Factorization Problem

Orthogonal nonnegative matrix factorization (ONMF) is an approximate matrix factorization technique with matrix orthogonality conditions and nonnegativity constraints, recently employed and shown to work remarkably well for clustering tasks such as document classification. Here, we briefly review some clustering related problems (k-means, ONMF and the isoperimetry) and describe their connections.Then, using the connections between ONMF and k-means and between k-means and isoperimetry problem, we propose an algorithm to solve the ONMF problem, and show its competitiveness by comparing our obtained results onrandomly generated matrices with the ones obtained by other algorithms.

کلیدواژه ها:

Orthogonal nonnegative matrix factorization ، clustering ، isoperimetry ، k-means