具自適應核形狀參數的徑向基底函數網路

Radial Basis Function Networks with Adaptive Kernel Shape Parameter

10.6382/JIM.200804.0007

葉怡成(I-Cheng Yeh)；陳重志(Chung-Chih Chen)；黃冠傑(Kuan-Chieh Huang)

半徑基神經網路 ； 監督式學習 ； 核函數 ； 分類 ； Radial basis function network ； supervised learning ； kernel function ； classification

資訊管理學報

15卷2期（2008 / 04 / 01）

135 - 154

繁體中文

徑向基底函數網路(RBFN)常用於分類問題，它的核有形心與半徑二種參數，這二種參數可用監督式或無監督式學習來決定。但它有一個缺點是視所有自變數有同等地位，故分類邊界是圓形，但事實上每一個自變數對分類的影響力不同，分類邊界是應該是橢圓形較合理。為克服此一缺點，本文提出具自適應核形狀參數的徑向基底函數網路，並以監督式學習推導出其學習規則。為證明此一架構優於傳統的徑向基底函數網路，本研究以五個人為的與七個真實的分類例題進行比較。結果顯示，此一架構確實比倒傳遞網路及傳統的徑向基底函數網路更為準確，狀參數值的大小確實能表現出自變數對分類的影響力高低。

Radial Basis Function Network (RBFN) is usually employed for classification problems, whose kernel has centroid and radius parameters determined with supervised or unsupervised learning. However, it has a shortcoming that it regards each independent variable as the same position; hence, the boundary of classification is circle. But in fact, each independent variable has different influence to the classification, it is more reasonable that the boundary of classification is ellipse. To overcome the shortcoming, we proposed the RBFN with adaptive kernel shape parameters and deduced its learning rule, using supervised learning. To verify whether the architecture is more accurate than conventional RBFN, experiments with five human-made problems and seven real-world problems were conducted. The results showed that this architecture is really more accurate than Back-Propagation Network and conventional RBFN, and the shape parameters can represent the influence of independent variable to classification.

1. 陳安斌、張志良(2001)。基因演算法自動演化之類神經網路在選擇權評價及避險之研究：分析與實證。資訊管理學報，7(2)，63-80。
   連結：
2. 陳安斌、許育嘉(2004)。整合小波轉換與神經網路於金融投資決策時間序列預測之研究。資訊管理學報，11(1)，139-165。
   連結：
3. 駱至中、林錦昌(2005)。尋優適應性類神經模糊推論模式於DRGs取巧行為自動檢測之應用。資訊管理學報，12(4)，53-74。
   連結：
4. Berthold, M.R.(1994).The TDRBF: A Shift Invariant Radial Basis Function Network.in Proc. of the Irish Neural Network Conference
5. Blackard, J. A.(1998).Fort Collins, Colorado,Department of Forest Sciences, Colorado State University.
6. Bugmann, G.(1998).Normalized Gaussian Radial Basis Function Networks.Neurocomputing (Special Issue on Radial Basis Function Networks),20(1),97-110.
7. Carse, B.,Pipe, A.G.,Fogarty, T.C.,Hill, T.(1995).Evolving Radial Basis Function Neural Networks Using a Genetic Algorithm.IEEE International Conference on Evolutionary Computation,Perth, WA, Australia:
8. Cheng, Y.M.(1997).Adaptive Rival Penalized Competitive Learning and Combined Linear Predictor Model for Financial Forecast and Investment.International Journal of Neural systems,8(5),517-534.
9. Gao, D.,Yang, G.(2002).Adaptive RBF Neural Networks for Pattern Classifications.Proc eedings of the 2002 International Joint Conference on Neural Networks
10. Gomm, J.B.,Dimg, L.Y.(2000).Selecting Radial Basis Function Network Centers with Recursive Orthogonal Least Squares Training.IEEE Transactions Neural Networks,11(2),306-314.
11. Han, M.,Xi, J.(2002).International Joint Conference on Neural Networks.Honolulu, HI:
12. PolyAnalyst Case Studies
13. Moody, J.,Darken, C.(1989).Fast Learning in Networks of Locally-Tuned Processing Units.Neural Computation,1(2),281-294.
14. Park, J.,Harley, R.G.,Venayagamoorthy, G.K.(2004).Indirect adaptive control for synchronous Generator: comparison of MLP/RBF neural networks approach with Lyapunov stability analysis.IEEE Transactions on Neural Networks,15(2),460-464.
15. Shibata, K.,Ito, K.(1999).International Joint Conference on Neural Networks.Washington, DC:
16. Statlog Project Databases
17. Webb, A.R.,Shannon, S.(1998).Shape-Adaptive Radial Basis Functions.IEEE Transactions on Neural Networks,9(6),1155-1166.
18. Whitehead, B. A.,Choate, T. D.(1996).Cooperative-Competitive Genetic Evolution of Radial Basis Function Centers and Widths for Time Series Prediction.IEEE Transactions on Neural Networks,7(4),869-880.
19. Xu, L.(1993).Rival Penalized Competitive Learning for Clustering Analysis RBF Net, and Curve Detection.IEEE Transactions on Neural Networks,4(4),636-648.
20. Yeh, I-Cheng(1999).Modeling Chaotic Two-Dimensional Mapping with Fuzzy-Neuron Networks.Fuzzy Sets and Systems,105(3),421-427.
21. 文少宣(2004)。碩士論文(碩士論文)。中華大學土木工程學系。
22. 葉怡成(2006)。類神經網路－方法應用與實作。台北:儒林書局。
23. 鄒明誠、孫志鴻(2004)。預測型模式在空間資料探勘之比較與整合研究。地理學報，38，93-109。