一個基於區間型Type-2模糊C-均值演算法之T-S模糊建模方法

A T-S Fuzzy Modeling Approach Based on Interval Type-2 Fuzzy C-means Algorithm

施毓洲(Yu-Zhou Shi)

指導教授 ： 蔡舜宏

國立臺北科技大學/機電學院/自動化科技研究所/碩士(2014年)

https://doi.org/10.6841/NTUT.2014.00748

摘要

本論文提出一個基於區間型Type-2模糊C-均值演算法之T-S模糊建模方法。為了建立模糊模型，第一步是確定模糊規則數，利用區間型Type-2模糊 C-均值(Interval Type-2 Fuzzy C-Means, IT2FCM)演算法將資料點做分群並且找到合適的分群數。基於Xie-Beni指數標準，將分群數目當作規則數，並且確定最優的規則數。接著利用模糊C-回歸模型(Fuzzy C-Regression Model, FCRM)演算法把未知系統分成數個線性系統，再依據數據資料的輸入與輸出，建立出各個線性系統的模糊規則參數初始值。經由過程所提供的參數訊息，並透過正交最小均方法得到線性系統最佳的參數值，T-S模糊(Takagi-Sugeno Fuzzy)模型即可建立出來。
最後，結合Type-2模糊C-均值演算法、Xie-Beni指數、模糊C-回歸模型的概念，透過一些例子，利用本論文的方法進行模糊建模，再來與其他方法做比較，經由實驗結果證明了本論文提出的建模方法可以比其他現有的方法更好。

關鍵字

T-S模糊模型 ； 模糊分群法 ； Type-2模糊 ； 參數辨識

並列摘要

This paper presents a T-S fuzzy modeling approach based on interval Type-2 fuzzy C-means algorithm. For fuzzy modeling, the first step is to determine the number of fuzzy rules. For this reason, fuzzy interval Type-2 C-means algorithm is adopted to classify the data points and determine the numbers of the cluster. Based on Xie-Beni index criterion, by defining the cluster numbers as the rule number, and then the optimal fuzzy rule number can be determined. Moreover, fuzzy c-regression model (FCRM) algorithm is adopted to divide unknown system into several linear systems. Based on the input and output data, we can establish the fuzzy rule parameters initial values of each linear system. According to the parameter messages from the process, the orthogonal least squares method can obtain the optimal parameter values for linear systems. Therefore, the Takagi-Sugeno fuzzy model can be established.
Finally, based on the conception of Type-2 fuzzy C-means algorithm, Xie-Beni index and FCRM, some examples are illustrated to demonstrate that the proposed modeling method can be better than the other existing methods.

並列關鍵字

T-S Fuzzy Model ； Fuzzy Clustering ； Type-2 Fuzzy ； Parameter Identify

參考文獻

[1] J. Q. Chen and L. J. Chen, "An on-line identification algorithm for fuzzy systems," Fuzzy Sets and Systems, vol. 64, no. 1, 1994, pp. 63-72.

[2] W. Pedrycz, "Identification algorithm in fuzzy relational systems," Fuzzy Sets and Systems, vol. 13, 1984, pp. 153-167.

[3] C. W. Xu and Y. Z. Lu, "Fuzzy model identification and self-learning for dynamic systems," IEEE Transactions on Man and Cybernetics Systems, vol. 17, no. 4, 1987, pp. 683-689.

[4] C. T. Sun, "Rule-based structure identification in an adaptive network based fuzzy inference system," IEEE Transactions on Fuzzy Systems, vol. 2, no. 1, 1994, pp. 64-73.

[5] S. Horikawa, T. Furuhashi and Y. Uchikawa, "On fuzzy modeling using fuzzy neural networks with the back-propagation algorithm," IEEE Transactions on Neural Networks, vol. 3, no. 5, 1992, pp. 801-806.