Title :
On the Feature Selection Criterion Based on an Approximation of Multidimensional Mutual Information
Author :
Balagani, Kiran S. ; Phoha, Vir V.
Author_Institution :
Center for Secure Cyberspace, Comput. Sci., Louisiana Tech Univ., Ruston, LA, USA
fDate :
7/1/2010 12:00:00 AM
Abstract :
We derive the feature selection criterion presented in [1] and [2] from the multidimensional mutual information between features and the class. Our derivation: 1) specifies and validates the lower-order dependency assumptions of the criterion and 2) mathematically justifies the utility of the criterion by relating it to Bayes classification error.
Keywords :
Bayes methods; pattern classification; Bayes classification error; feature selection criterion; lower-order dependency assumption; multidimensional mutual information; Bayes classification error; Feature selection; entropy; entropy estimation.; mutual information;
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on
DOI :
10.1109/TPAMI.2010.62
