基于递归聚类与相似性的模糊神经网络结构设计

李微; 乔俊飞

doi:10.11936/bjutxb2016040086

基于递归聚类与相似性的模糊神经网络结构设计

doi: 10.11936/bjutxb2016040086

李微^1,,
乔俊飞¹

1.
北京工业大学信息学部, 北京 100124
2.
计算智能与智能系统北京市重点实验室, 北京 100124

基金项目: 国家自然科学基金资助项目(61533002);国家杰出青年科学基金资助项目(61225016);北京市科技新星计划(Z131104000413007)

详细信息

中图分类号: TP183
计量
- 文章访问数: 99
- HTML全文浏览量: 56
- PDF下载量: 0
- 被引次数: 0
出版历程
- 收稿日期: 2016-04-26
- 网络出版日期: 2022-09-13
- 刊出日期: 2017-02-01

Structure Design of Fuzzy Neural Networks Based on Recursive Clustering and Similarity

LI Wei^1
,,
QIAO Junfei¹

1.
Faculty of Information Technology, Beijing University of Technology, Beijing 100124, China
2.
Beijing Key Laboratory of Computational Intelligence and Intelligent System, Beijing 100124, China

摘要

摘要: 针对模糊神经网络结构设计问题,提出一种基于递归聚类与相似性的结构设计方法. 首先,提出以输出变化强度为导向、以结构细分为手段的递归聚类方法对网络初始结构进行设计. 其次,通过计算模糊规则的相似性,将高度相似的规则进行合并,在保持良好精度的前提下,对网络初始结构进行简化. 最后,通过函数逼近、非线性系统辨识仿真实验验证了方法的可行性和有效性.
- 模糊神经网络 /
- 结构设计 /
- 递归聚类 /
- 相似性
Abstract: Facing the structure design problem of fuzzy neural networks (FNNs), this paper proposed a structure design approach based on the recursive clustering and similarity methods. First, a recursive clustering method to identify FNN structure was proposed. Guided by the strength of output variations and using the recursive sub-clustering as the means, the proposed method determined the initial network structure through recursive iterations. Second, maintaining a high accuracy, the method calculated the similarity degree between each pair of fuzzy rules and then merged highly similar rules to simplify the initialized structure of the FNN. Finally, numerical experiments in function approximation and nonlinear system identification were used to verify the feasibility and effectiveness of the proposed approach.
- fuzzy neural networks /
- structure design /
- recursive clustering /
- similarity
The authors have declared that no competing interests exist.

HTML全文

图 1 Mamdani模糊神经网络逼近效果

Figure 1. Approximation result of Mamdani FNN

下载: 全尺寸图片幻灯片

图 2 Mamdani模糊神经网络逼近误差

Figure 2. Approximation error of Mamdani FNN

下载: 全尺寸图片幻灯片

图 3 Mamdani模糊神经网络辨识效果

Figure 3. Identification result of Mamdani FNN

下载: 全尺寸图片幻灯片

图 4 Mamdani模糊神经网络辨识误差

Figure 4. Identification error of Mamdani FNN

下载: 全尺寸图片幻灯片

表 1 不同方法的性能对比

Table 1. Comparisons between different methods

方法	模糊规则数	参数个数	训练RMSE	测试RMSE
Wang等^[8]	8	24	0.0241	0.0492
Pedrycz等^[7]	8	24	0.150	0.153
Pedrycz等^[7]	10	30	0.1470	0.1620
Wang等^[15]	11	33	0.0018	0.0225
本文	8	24	0.0167	0.0395
本文	10	30	0.0100	0.0168

下载: 导出CSV

表 2 不同方法的性能对比

Table 2. Comparisons between different methods

方法	模糊规则数	参数个数	测试RMSE
DFNN^[14]	6	48	0.1315
SOFMLS^[17]	6	48	0.0290
GA^[18]	7	37	0.0500
MCFC^[19]			0.0874
本文(结构简化前)	6	42	0.0207
本文(结构简化后)	5	35	0.0259

下载: 导出CSV

参考文献(19)

[1]	KUNG C C, SU J Y.Affine Takagi-Sugeno fuzzy modelling algorithm by fuzzy c-regression models clustering with a novel cluster validity criterion[J]. IET Control Theory and Application, 2007, 1(5): 1255-1265.
[2]	LI C S, ZHOU J Z, FU B, et al.T-S fuzzy model identification with a gravitational search-based hyperplane clustering algorithm[J]. IEEE Transactions on Fuzzy System, 2012, 20(2): 305-317.
[3]	PEDRYCZ W, IZAKIAN H.Cluster-centric fuzzy modeling[J]. IEEE Transactions on Fuzzy System, 2014, 22(6): 1585-1597.
[4]	CHEN M Y, LINKENS D A.Rule-base self-generation and simplification for data-driven fuzzy models[J]. Fuzzy Sets and Systems, 2004, 142(2): 243-265.
[5]	DELGADO M, GOMEZ-SKARMETA A F, MARTIN F. A fuzzy clustering-based rapid prototyping for fuzzy rule-based modeling[J]. IEEE Transactions on Fuzzy System, 1997, 5(2): 223-233.
[6]	PRASAD M, LIN Y Y, LIN C Y, et al.A new data-driven neural fuzzy system with collaborative fuzzy clustering mechanism[J]. Neurocomputing, 2015, 167: 558-568.
[7]	PEDRYCZ W, KWAK K C.Linguistic models as a framework of user-centric system modeling[J]. IEEE Transactions on Systems, Man, and Cybernetics—Part A: Systems and Humans, 2006, 36(4): 727-745.
[8]	WANG D, ZENG X J, KEANE J A.An output-constrained clustering approach for the identification of fuzzy systems and fuzzy granular systems[J]. IEEE Transactions on Fuzzy System, 2011, 19(6): 1127-1140.
[9]	CHAO C T, CHEN Y J, TENG C C.Simplification of fuzzy neural systems using similarity analysis[J]. IEEE Transactions on Systems, Man, and Cybernetics—Part B: Cybernetics, 1996, 26(2): 344-354.
[10]	MACQUEEN J.Some methods for classification and analysis of multivativate observations [C]∥Proceedings of 5-th Berkeley Symposium on Mathematical Statistics and Probability. Berkeley: University of California Press, 1967: 281-297.
[11]	李国勇. 神经模糊控制理论及应用[M]. 北京: 电子工业出版社, 2009: 286-316.
[12]	LENG G, PRASAD G, MCGINNITY T M.An on-line algorithm for creating self-organizing fuzzy neural networks[J]. Neural Networks, 2004, 17(10): 1477-1493.
[13]	SETNES M, BABUSKA R, KAYMAK U, et al.Similarity measures in fuzzy rule base simplification[J]. IEEE Transactions on Systems, Man, and Cybernetics—Part B: Cybernetics, 1998, 28(3): 376-386.
[14]	LI W, QIAO J F, ZENG X J.Accurate similarity analysis and computing of Gaussian membership functions for FNN simplification[C]//12th International Conference on Fuzzy Systems and Knowledge Discovery. New York: Institute of Electrical and Electronics Engineers Inc, 2015: 402-409.
[15]	WANG D, ZENG X J, KEANE J A.A clustering algorithm for radial basis function neural network initialization[J]. Neurocomputing, 2012, 77: 144-155.
[16]	ER M J, WU S Q.A fast learning algorithm for parsimonious fuzzy neural systems[J]. Fuzzy Sets and Systems, 2002, 126(3): 337-351.
[17]	RUBIO J D J. SOFMLS: online self-organizing fuzzy modified least-squares network[J]. IEEE Transaction on Fuzzy Systems, 2009, 17(6): 1296-1309.
[18]	GUENOUNOUA O, DAHHOU B, CHABOUR F.TSK fuzzy model with minimal parameters[J]. Applied Soft Computing, 2015, 30: 748-757.
[19]	PRASADMCHOU KPSAXENAAet al.Collaborative fuzzy rule learning for Mamdani type fuzzy inference system with mapping of cluster centersProceedings of 2014 IEEE Symposium Series on Computational Intelligence. New York: Institute of Electrical and Electronics Engineers Inc20141520 PRASAD M, CHOU K P, SAXENA A, et al.Collaborative fuzzy rule learning for Mamdani type fuzzy inference system with mapping of cluster centers[C]//Proceedings of 2014 IEEE Symposium Series on Computational Intelligence. New York: Institute of Electrical and Electronics Engineers Inc, 2014: 15-20.