Medical diagnosis applications using a novel interactively recurrent self-evolving fuzzy CMAC model

• Published in: 2014 International Joint Conference on Neural Networks (IJCNN) pp. 4092 - 4098
• Main Authors: Jyun-Guo Wang, Shen-Chuan Tai, Cheng-Jian Lin
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.07.2014
• Subjects: Computational modeling; Firing; gradient descent algorithm; Hypercubes; Input variables; interactively recurrent self-evolving fuzzy Cerebellar Model Articulation Controller; Mathematical model; Training data; Vectors
• ISSN: 2161-4393
• DOI: 10.1109/IJCNN.2014.6889511

In this paper, a recurrent self-evolving Fuzzy Cerebellar Model Articulation Controller (FCMAC) model for classification problems is developed, namely the interactively recurrent self-evolving fuzzy Cerebellar Model Articulation Controller (IRSFCMAC). The interactively recurrent structure in an IRSFCMAC is formed as external loops and internal feedbacks by feeding the rule firing strength to itself and others rules. The IRSFCMAC learning starts with an empty rule base and all of rules are generated and learned online, through a simultaneous structure and parameter learning, while the relative parameters are learned through a gradient descent algorithm. The proposed IRSFCMAC is tested by the four benchmarked classification problems and compared with the well-known traditional FCMAC. Experimental results show that the proposed IRSFCMAC model enhanced classification performance results, in terms of accuracy and RMSE.