J. Comp. Math., 30 (2012), pp. 34-46.


An Effective Initialization for Orthogonal Nonnegative Matrix Factorization

Xuansheng Wang 1, Xiaoyao Xie 2, Linzhang Lu 3

1 School of Mathematical Science, Xiamen University, Xiamen 361005, China
2 School of Mathematics and Computer Science, Guizhou Normal University, Guiyang 550001, China
3 School of Mathematics and Computer Science, Guizhou Normal University, Guiyang 550001, China
3 School of Mathematical Science, Xiamen University, Xiamen 361005, China

Received 2011-2-28 Accepted 2011-6-20
Available online 2012-1-09
doi:10.4208/jcm.1110-m11si10

Abstract

The orthogonal nonnegative matrix factorization (ONMF) has many applications in a variety of areas such as data mining, information processing and pattern recognition. In this paper, we propose a novel initialization method for the ONMF based on the Lanczos bidiagonalization and the nonnegative approximation of rank one matrix. Numerical experiments are given to show that our initialization strategy is effective and efficient.

Key words: Lanczos bidiagonalization, Orthogonal nonnegative matrix factorization, Low-rank approximation, Nonnegative approximation.

AMS subject classifications: 65F99.


Email: wxs111111@163.com, xyx@gznu.edu.cn, lzlu@xmu.edu.cn
 

The Global Science Journal