A Hybrid RBF Network based on Fuzzy Dynamic Learning Rate Control

Kim， Kwang-baek; Park Choong Shik

Journal of The Korea Society of Computer and Information

Abbr : JKSCI
2014, 19(9), pp.33-38
Publisher : The Korean Society Of Computer And Information
Research Area : Computer Science

Kim， Kwang-baek ¹, Park Choong Shik ²

¹신라대학교
²영동대학교

ABSTRACT

The FCM based hybrid RBF network is a heterogeneous learning network model that appliesFCM algorithm between input and middle layer and applies Max_Min algorithm between middlelayer and output. The Max-Min neural network uses winner nodes of the middle layer as input butshows inefficient learning in performance when the input vector consists of too many patterns. To overcome this problem, we propose a dynamic learning rate control based on fuzzy logic. Theproposed method first classifies accurate/inaccurate class with respect to the difference betweentarget value and output value with threshold and then fuzzy membership function and fuzzydecision logic is designed to control the learning rate dynamically. We apply this proposed RBFnetwork to the character recognition problem and the efficacy of the proposed method is verifiedin the experiment.
KEYWORDS

FCMbased hybrid network, Max-Min neural network, Fuzzy logic, Learning rate
REFERENCES (9)

[book] D. K. Arun / 2001 / Computer Vision and Fuzzy-Neural Systems / Prentice Hall PTR

[journal] D. Q Zhang / 2003 / Clustering Incomplete Data Using Kernel-Based Fuzzy C-means Algorithm / Neural Processing Letters 18(3) : 155~162

[confproc] L. S. Hsu / 1992 / Fuzzy Neural-Logic System / Proceedings of IJCNN 1 : 254~250

[confproc] A. Blanco / 1994 / Solving fuzzy relational equations by max-min neural networks / Proceedings of IEEE World Congress on Computational Intelligence 3 : 1737~1742

[confproc] K. B. Kim / 1997 / A Fuzzy Competitive Backpropagation using Nervous System / Proceedings of World Congress on Systems Simulation : 188~193

[journal] K. B. Kim / 2005 / An Enhanced Fuzzy Neural Network / Lecture Notes in Computer Science 3320 : 176~179

[journal] K. B. Kim / 2005 / Hierarchical Recognition of English Calling Card by Using Multiresolution Images and Enhanced Neural Network / Lecture Notes in Computer Science 3801 : 785~792

[journal] K. B. Kim / 2013 / Enhanced FCM-based Hybrid Network / INFORMATION-AN INTERNATIONAL INTERDISCIPLINARY JOURNAL 16(2B) : 1509~1515

[journal] K. B. Kim / 2004 / Performance Improvement of RBF Network using ART2 Algorithm and Fuzzy Logic System / Lecture Notes in Artificial Intelligence, LNAL / Springer 3339 : 853~860
KCI Citation (0)

KJC Korea
Journal Central

Statistics

Tools