||Improving Non-negative Matrix Factorizations Through Structured Initialization
||In this paper we explore a recent iterative compression technique called non-negative matrix factorization (NMF). Several
special properties are obtained as a result of the constrained optimization problem of NMF. For facial images, the additive
nature of NMF results in a basis of features, such as eyes, noses, and lips. We explore various methods for e4ciently computing
NMF, placing particular emphasis on the initialization of current algorithms. We propose using Spherical K-Means clustering
to produce a structured initialization for NMF. We demonstrate some of the properties that result from this initialization and
develop an e4cient way of choosing the rank of the low-dimensional NMF representation.
||Non-negative matrix factorization, K-means clustering, Constrained optimization, Rank reduction, Data mining, Compression, Feature extraction
||This work initiated under NSF VIGRE Grant No. DMS-9810751
||Appears in Pattern Recognition, Volume 37, Issue 11, November 2004, Pages 2217-2232.
||[PDF via Science Direct]|
author = "Stefan M. Wild and James H. Curry and Anne Dougherty",
title = "Improving Non-negative Matrix Factorizations Through Structured Initialization",
journal = "Pattern Recognition",
volume = "37",
number = "11",
month = "November",
year = "2004",
pages = "2217--2232",
doi = "http://dx.doi.org/10.1016/j.patcog.2004.02.013"