Summary
International Symposium on Nonlinear Theory and its Applications
2009
Session Number:B3L-C
Session:
Number:B3L-C4
An efficient PCA based method for Tucker3 model selection
Zhaoshui He, Andrzej Cichocki,
pp.-
Publication Date:2009/10/18
Online ISSN:2188-5079
Summary:
There has been a growing interest in Tucker3 analysis recently. One of the biggest challenges in Tucker3 analysis is the model selection problem: how to choose the number of components in each mode of an observed tensor. An alternative Tucker3 model selection approach is developed based on principal component analysis (PCA) for this problem in this paper. It is computationally efficient and straightforward to implement. Its effectiveness is demonstrated by the experiments.