DocumentCode
458929
Title
New developments of the Z-EDM algorithm
Author
Silva, Lauis M. ; De Sa, J. Marques ; Alexandre, Luis A.
Author_Institution
Instituto de Engenharia Biomedica, Porto
Volume
1
fYear
2006
fDate
16-18 Oct. 2006
Firstpage
1067
Lastpage
1072
Abstract
In this paper we address some open questions on the recently proposed Zero-Error Density Maximization algorithm for MLP training. We propose a new version of the cost function that solves a training problem encountered in previous work and prove that the use of a nonparametric density estimator preserves the optimal solution. Some experiments are reported comparing this cost function to the usual mean-square error and cross entropy cost functions
Keywords
learning (artificial intelligence); multilayer perceptrons; optimisation; parameter estimation; MLP training; Zero-Error Density Maximization algorithm; cross entropy cost functions; mean-square error; nonparametric density estimator; Cost function; Encoding; Entropy; Error correction; Hypercubes; Intelligent systems; Kernel; Multilayer perceptrons; Pattern recognition; Smoothing methods;
fLanguage
English
Publisher
ieee
Conference_Titel
Intelligent Systems Design and Applications, 2006. ISDA '06. Sixth International Conference on
Conference_Location
Jinan
Print_ISBN
0-7695-2528-8
Type
conf
DOI
10.1109/ISDA.2006.204
Filename
4021587
Link To Document