Efficient Self-Adaptive Learning Algorithm for TSK-Type Compensatory Neural Fuzzy Networks

In this paper, a TSK-type compensatory neural fuzzy network (TCNFN) for classification applications is proposed. The TCNFN model is a five-layer structure, which combines the traditional Takagi-Sugeno-Kang (TSK). Layer 3 of the TCNFN model contains adaptive compensatory fuzzy operations, which make fuzzy logic systems more adaptive and effective. Furthermore, a self-adaptive learning algorithm, which consists of the structure learning and the parameter learning, is also proposed. The structure learning is based on the degree measure to determine the number of fuzzy rules and the parameter learning is based on the gradient descent algorithm to adjust the parameters of the TCNFN. The advantages of the proposed method are that it converges quickly and that the fuzzy rules that are obtained are more precise. The performance of TCNFN compares excellently with other various existing methods.

[1]  Nikola K. Kasabov,et al.  Evolving fuzzy neural networks for supervised/unsupervised online knowledge-based learning , 2001, IEEE Trans. Syst. Man Cybern. Part B.

[2]  P. K. Simpson.  Fuzzy Min-Max Neural Networks-Part 1 : Classification , 1992 .

[3]  Chen-Chia Chuang,et al.  Two-Stages Support Vector Regression for Fuzzy Neural Networks with Outliers , 2009 .

[4]  Chin-Teng Lin,et al.  An online self-constructing neural fuzzy inference network and its applications , 1998, IEEE Trans. Fuzzy Syst..

[5]  Chia-Feng Juang,et al.  A Self-Organizing TS-Type Fuzzy Network With Support Vector Learning and its Application to Classification Problems , 2007, IEEE Transactions on Fuzzy Systems.

[6]  Chih-Ming Chen,et al.  An efficient fuzzy classifier with feature selection based on fuzzy entropy , 2001, IEEE Trans. Syst. Man Cybern. Part B.

[7]  Rudolf Kruse,et al.  A neuro-fuzzy method to learn fuzzy classification rules from data , 1997, Fuzzy Sets Syst..

[8]  Richard O. Duda,et al.  Pattern classification and scene analysis , 1974, A Wiley-Interscience publication.

[9]  Shyi-Ming Chen,et al.  A new method for constructing membership functions and fuzzy rules from training examples , 1999, IEEE Trans. Syst. Man Cybern. Part B.

[10]  Brian C. Lovell,et al.  The Multiscale Classifier , 1996, IEEE Trans. Pattern Anal. Mach. Intell..

[11]  Satish Kumar,et al.  Subsethood-product fuzzy neural inference system (SuPFuNIS) , 2002, IEEE Trans. Neural Networks.

[12]  Huseyin Seker,et al.  Compensatory fuzzy neural networks-based intelligent detection of abnormal neonatal cerebral Doppler ultrasound waveforms , 2001, IEEE Transactions on Information Technology in Biomedicine.

[13]  Patrick K. Simpson,et al.  Fuzzy min-max neural networks. I. Classification , 1992, IEEE Trans. Neural Networks.

[14]  Jeen-Shing Wang,et al.  Self-adaptive neuro-fuzzy inference systems for classification applications , 2002, IEEE Trans. Fuzzy Syst..

[15]  Sushmita Mitra,et al.  Neuro-fuzzy rule generation: survey in soft computing framework , 2000, IEEE Trans. Neural Networks Learn. Syst..

[16]  Marco Russo,et al.  FuGeNeSys-a fuzzy genetic neural system for fuzzy modeling , 1998, IEEE Trans. Fuzzy Syst..

[17]  Stanislaw Osowski,et al.  ECG beat recognition using fuzzy hybrid neural network , 2001, IEEE Trans. Biomed. Eng..

[18]  Abraham Kandel,et al.  Compensatory neurofuzzy systems with fast learning algorithms , 1998, IEEE Trans. Neural Networks.

[19]  Hahn-Ming Lee,et al.  A neural network classifier with disjunctive fuzzy information , 1998, Neural Networks.

[20]  Wei-Yen Wang,et al.  An on-line robust and adaptive T-S fuzzy-neural controller for more general unknown systems , 2008 .

[21]  Huan Liu,et al.  Neural-network feature selector , 1997, IEEE Trans. Neural Networks.