Fuzzy Relation-Based Polynomial Neural Networks Based on Hybrid Optimization

详细信息    查看全文

• 作者：Wei Huang (1)
  Sung-Kwun Oh (2) ohsk@suwon.ac.kr
• 关键词：Hybrid optimized fuzzy relation ; based polynomial neural network (HOFRPNN) – ; hybrid optimization – ; fuzzy rule ; based models – ; polynomial neural networks (PNNs)
• 刊名：Lecture Notes in Computer Science
• 出版年：2012
• 出版时间：2012
• 年：2012
• 卷：7367
• 期：1
• 页码：90-97
• 全文大小：204.9 KB
• 参考文献：1. Sommer, V., Tobias, P., Kohl, D., Sundgren, H., Lundstrom, L.: Neural Networks and Abductive Networks for Chemical Sensor Signals: A Case Comparison. Sens. Actuators B, Chem. 28, 217–222 (1995)
  2. Cherkassky, V., Gehring, D., Mulier, F.: Comparison of Adaptive Methods for Function Estimation from Samples. IEEE Transactions on Neural Networks 7, 969–984 (1996)
  3. Woei, W.T., Hong, H.: A Generic Neurofuzzy Model-Based Approach for Detecting Faults in Induction Motors. IEEE Trans. Ind. Electron. 52, 1478–1489 (2005)
  4. Ivakhnenko, A.G.: Polynomial Theory of Complex Systems. IEEE Trans. Syst., Man, Cybern. B, Cybern. SMC 1, 364–378 (1971)
  5. Ohtani, T., Ichihashi, H., Miyoshi, T., Nagasaka, K.: Orthogonal and Successive Projection Methods for the Learning of Neurofuzzy GMDH. Information Sciences 110, 5–24 (1998)
  6. Oh, S.K., Pderycz, W.: The Design of Self-Organizing Polynomial Neural Networks. Information Sciences 141, 237–258 (2002)
  7. Park, B.J., Pedrycz, W., Oh, S.K.: Fuzzy Polynomial Neural Networks: Hybrid Architectures of Fuzzy Modeling. IEEE Transactions on Fuzzy Systems 10, 607–621 (2002)
  8. Huang, W., Ding, L., Oh, S.K., Jeong, C.W., Joo, S.C.: Identification of Fuzzy Inference System Based on Information Granulation. KSII Transactions on Internet and Information Systems 4, 575–593 (2010)
  9. Pedrycz, W.: An Identification Algorithm in Fuzzy Relational System. Fuzzy Sets Syst. 13, 153–167 (1984)
  10. Tong, R.M.: The Evaluation of Fuzzy Models Derived from Experimental Data. Fuzzy Sets Syst. 13, 1–12 (1980)
  11. Xu, C.W., Zailu, Y.: Fuzzy Model Identification Self-learning for Dynamic System. IEEE Trans. Syst., Man, Cybern. B, Cybern. 17, 683–689 (1987)
  12. Sugeno, M., Yasukawa, T.: Linguistic Modeling Based on Numerical Data. In: IFSA 1991 Brussels, Computer, Management & System Science, pp. 264–267. IEEE Press (1991)
  13. Oh, S.K., Pedrycz, W.: Identification of Fuzzy Systems By means of An Auto-Tuning Algorithm and Its Application to Nonlinear Systems. Fuzzy Sets and Syst. 115, 205–230 (2000)
  14. Park, B.J., Pedrycz, W., Oh, S.K.: Identification of Fuzzy Models with the Aid of Evolutionary Data Granulation. IEE Proc. Control Theory and Applications 148, 406–418 (2001)
  15. Pedrycz, W., Park, H.S., Oh, S.K.: A Granular-Oriented Development of Functional Radial Basis Function Neural Networks. Neurocomputing 72, 420–435 (2008)
  16. Pedrycz, W., Kwak, K.C.: Linguistic models as a framework of user-centric system modeling. IEEE Trans. Syst., Man Cybern. –PART A: Systems and Humans 36, 727–745 (2006)
• 作者单位：1. School of Computer and Communication Engineering, Tianjin University of Technology, Tianjin, 300191 China2. Department of Electrical Engineering, The University of Suwon, San 2-2, Wau-ri, Bongdam-eup, Hwaseong-si, Gyeonggi-do 445-743, South Korea
• ISSN：1611-3349

文摘

This paper introduces hybrid optimized fuzzy relation-based polynomial neural network (HOFRPNN), a novel architecture that is constructed by using a combination of fuzzy rule-based models, polynomial neural networks (PNNs) and a hybrid optimization algorithm. The proposed hybrid optimization algorithm is developed by a combination of a space search algorithm and an improved complex method. The structure of HOFRPNN comprises of a synergistic usage of fuzzy-rule-based polynomial neuron that are essentially fuzzy rule-based models and polynomial neural networks that is an extended group method of data handling (GMDH). The architecture of HOFRPNN is an essentially modified PNN whose basic nodes are fuzzy-rule-based polynomial neurons rather than conventional polynomial neurons. Moreover, the hybrid optimization algorithm is utilized to optimize the structure topology of HOFRPNN. A comparative study demonstrates that the proposed model exhibits higher accuracy and superb predictive capability when compared with some previous models reported in the literature.