A Hybrid Bayesian Optimal Classifier Based on Neuro-fuzzy Logic

  • Hongsheng Su
  • Qunzhan Li
  • Jianwu Dang
Conference paper

DOI: 10.1007/11881070_49

Part of the Lecture Notes in Computer Science book series (LNCS, volume 4221)
Cite this paper as:
Su H., Li Q., Dang J. (2006) A Hybrid Bayesian Optimal Classifier Based on Neuro-fuzzy Logic. In: Jiao L., Wang L., Gao X., Liu J., Wu F. (eds) Advances in Natural Computation. ICNC 2006. Lecture Notes in Computer Science, vol 4221. Springer, Berlin, Heidelberg

Abstract

Based on neural networks and fuzzy set theory, a hybrid Bayesian optimal classifier is proposed in the paper. It can implement fuzzy operation, and generate learning behaviour. The model firstly applies fuzzy membership function of the observed information to establish the posterior probabilities of original assumptions in Bayesian classification space, the classified results of all input information then are worked out. Across the calculation, the positive and reverse instances of all observed information are fully considered. The best classification result is acquired by incorporating with all possible classification results. The whole classifier adopts a hybrid four-layer forward neural network to implement. Fuzzy operations of input information are performed using fuzzy logic neurons. The investigation indicates that the proposed method expands the application scope and classification precision of Bayesian optimal classifier, and is an ideal patter classifier. In the end, an experiment in transformer insulation fault diagnosis shows the effectiveness of the method.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer-Verlag Berlin Heidelberg 2006

Authors and Affiliations

  • Hongsheng Su
    • 1
  • Qunzhan Li
    • 1
  • Jianwu Dang
    • 2
  1. 1.School of Electrical EngineeringSouthwest Jiaotong UniversityChengduChina
  2. 2.School of Information and Electrical EngineeringLanzhou Jiaotong UniversityLanzhouChina

Personalised recommendations