Title :
A novel neuro-fuzzy classification system design by a species-based hybrid algorithm
Author :
Lee, Ching-Hung ; Chiu, Hsin-Wei ; Li, Chung-Ta
Author_Institution :
Dept. of Electr. Eng., Yuan-Ze Univ., Taoyuan, Taiwan
Abstract :
In this paper, we propose a novel neuro-fuzzy classification system by a species-based hybrid of electromagnetism-like mechanism and back-propagation algorithms (SEMBP). The neuro-fuzzy classification system is constructed by an interval type-2 fuzzy neural system with asymmetric membership functions (AIT2FNS). The hybrid algorithm SEMBP combines the advantages of EM and BP algorithms. Three classification problems: the XOR data set, the breast cancer data set, and the Iris data set are used to illustrate the performance of our approach.
Keywords :
backpropagation; fuzzy neural nets; pattern classification; Iris data set; XOR data set; asymmetric membership function; backpropagation algorithm; breast cancer data set; electromagnetism like mechanism; hybrid algorithm SEMBP; interval type-2 fuzzy neural system; neurofuzzy classification system design; species based hybrid algorithm; Accuracy; Artificial neural networks; Breast cancer; Classification algorithms; Machine learning; Support vector machines; Training; Classification; TSK type; asymmetric membership function; type-2 fuzzy neural system; uniform initialization;
Conference_Titel :
Machine Learning and Cybernetics (ICMLC), 2010 International Conference on
Conference_Location :
Qingdao
Print_ISBN :
978-1-4244-6526-2
DOI :
10.1109/ICMLC.2010.5580807
