Adelaide Research and Scholarship : Schools and Disciplines : School of Civil, Environmental and Mining Engineering : Civil and Environmental Engineering Publications

Type:	Conference paper
Title:	Critical values of a kernel density-based mutual information estimator
Author:	May, Robert James Dandy, Graeme C. Maier, Holger R. Fernando, Thudugala Mudalige Kaushalya Gayani
Citation:	International Joint Conference on Neural Networks, 16-21 July 2006:pp.4898-4903
Publisher:	IEEE
Issue Date:	2006
ISBN:	0780394909
Conference Name:	International Joint Conference on Neural Networks (2006 : Vancouver, Canada) IEEE World Congress on Computational Intelligence (2006 : Vancouver, Canada) IJCNN '06
School/Discipline:	School of Civil and Environmental Engineering
Abstract:	Recently, mutual information (MI) has become widely recognized as a statistical measure of dependence that is suitable for applications where data are non-Gaussian, or where the dependency between variables is non-linear. However, a significant disadvantage of this measure is the inability to define an analytical expression for the distribution of MI estimators, which are based upon a finite dataset. This paper deals specifically with a popular kernel density based estimator, for which the distribution is determined empirically using Monte Carlo simulation. The application of the critical values of MI derived from this distribution to a test for independence is demonstrated within the context of a benchmark input variable selection problem.
Description:	Copyright © 2006 IEEE
RMID:	0020062794
DOI:	10.1109/IJCNN.2006.247170
Description (link):	http://www.okstate.edu/elec-engr/faculty/yen/wcci/WCCI-Web_ProgramList_F.html
Links to content (authorised users):	Check full text options http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=1716780
Appears in Collections:	Civil and Environmental Engineering Publications
View citing articles in:	Google Scholar Scopus

Unless expressly stated otherwise, all items in Adelaide Research and Scholarship are protected by copyright, with all rights reserved.