A STUDY OF SEVERAL MODEL SELECTION CRITERIA FOR DETERMINING THE NUMBER OF SIGNALS
Refereed conference paper presented and published in conference proceedings


Times Cited
Web of Science1WOS source URL (as at 19/01/2021) Click here for the latest count
Altmetrics Information
.

Other information
AbstractAddressing the problem of detecting the number of source signals as selecting the hidden dimensionality of Factor Analysis (FA) model, we investigate several model selection criteria via a new empirical analyzing tool that examines the joint effect of signal-noise ratio (SNR) and sample size N on the model selection performance. The contours of the model selection accuracies visualize a three-region partition on the space of SNR and N, and a diminishing marginal effect which trades off SNR and N on the performance. Moreover, the newly derived Variational Bayes algorithm and three variants of Bayesian Ying-Yang (BYY) algorithms are more robust against reducing SNR and N, where the BYY with priors' hyperparameters updated is the best in general.
All Author(s) ListTu SK, Xu L
Name of Conference2010 IEEE International Conference on Acoustics, Speech, and Signal Processing
Start Date of Conference10/03/2010
End Date of Conference19/03/2010
Place of ConferenceDallas
Country/Region of ConferenceUnited States of America
Detailed descriptionorganized by IEEE,
Year2010
Month1
Day1
PublisherIEEE
Pages1966 - 1969
eISBN978-1-4244-4296-6
ISSN1520-6149
LanguagesEnglish-United Kingdom
Keywordscriteria; hidden dimensionality; linear model; model selection; Number of signals
Web of Science Subject CategoriesAcoustics; Computer Science; Computer Science, Theory & Methods; Engineering; Engineering, Electrical & Electronic

Last updated on 2021-20-01 at 00:58