研究生: |
孫士宸 Shih-Chen Sun |
---|---|
論文名稱: |
一個用於圖型辨識的指導式學習之模糊權重自適應共振網路 A Supervised Learning ART Network with Fuzzy Weight Adjustment for Pattern Recognition |
指導教授: |
楊英魁
Ying-Kuei Yang |
口試委員: |
蘇仲鵬
none 連耀南 none 孫宗瀛 none 吳傳嘉 Chwan-Chia Wu |
學位類別: |
碩士 Master |
系所名稱: |
電資學院 - 電機工程系 Department of Electrical Engineering |
論文出版年: | 2006 |
畢業學年度: | 94 |
語文別: | 中文 |
論文頁數: | 61 |
中文關鍵詞: | 自適應共振網路 、標準樣本 、非完整樣本 、分類辨識碼 |
外文關鍵詞: | ART neural network, standard patterns, incomplete patterns, recognition codes |
相關次數: | 點閱:168 下載:2 |
分享至: |
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
本論文是用來改善自適應共振網路(ART neural network)對非完整樣本的辨識能力。由於自適應共振網路為一非指導式、競爭式學習的類神經網路模型,使得非完整樣本容易對訓練效果造成負面影響。
自適應共振理論(Adaptive Resonance Theory, ART)對「穩定性—可塑性」的兩難問題(stability-plasticity dilemma)提供了一個解決方法。但是非指導式學習的方式無法分辨訓練樣本的種類,將可能降低了分類辨識碼的精確度。
本論文的方法是改良自適應共振網路,以指導式學習的方式,使自適應共振網路對不同的訓練樣本有不同的學習反應,並使用歸屬函數對權重做模糊調整。改良之後的自適應共振網路能對標準樣本產生精確的分類辨識碼來做記憶,而對非完整樣本做模糊的關連性學習,達到對非完整樣本進行辨識的功能,並藉由模糊權重調整提高辨識的準確度與速度。由於以自適應共振網路為基礎,能解決「穩定性¬—可塑性」的兩難問題。
根據模擬結果,本論文所提的方法對非完整樣本有快速且準確的學習與辨識能力。
This thesis proposes an enhanced adaptive resonance theory (ART) neural network to improve the capability of recognizing incomplete patterns.
ART provides a solution to the stability-plasticity dilemma. Nonetheless, the unsupervised learning algorithm can not distinguish standard patterns from the incomplete patterns during learning stage due to its unsupervised and competitive learning nature, which greatly degrades the accuracy rate of recognition.
The core ideas of the proposed approach in this thesis are: (1) Enhancing the ART neural network by supervised learning algorithm to create the capability of accepting both complete and incomplete learning patterns; and (2) Applying the concept of membership function in fuzzy theory to weight adjustment for network nodes to increase the accuracy rate of recognition. The enhanced ART is able to not only precisely memorize the classification codes of standard patterns but also learn fuzzy relationships for incomplete patterns.
The simulation results in this paper has shown the enhanced ART is able to learn and recognize incomplete patterns efficiently and correctly
[1] G.A. Carpenter and S. Grossberg, “A Massively Parallel Architecture for a Self-Organizing Neural Pattern Recognition Machine,” Computer Vision, Graphics, and Image Processing, Vol.37, pp.54-115(1987).
[2] G.A. Carpenter and S. Grossberg, “The ART of Adaptive Pattern Recognition by a Self-Organizing Neural Network,” Computer, Vol.21, No.3, pp.77-88(1988).
[3] A.P. Dempster, N.M. Laird and D.B. Rubin, “Maximum-likelihood from incomplete data via the EM algorithm,” J. R. Stat. Soc., Vol.B39, pp.1–38(1977).
[4] G.J. McLachlan and T. Krishnan, The EM Algorithm and Extensions, Wiley, New York(1997).
[5] Z. Ghahramani and M.I. Jordan, “Supervised Learning from Incomplete Data via an EM Approach,” Proc. Advances in Neural Information Processing Systems 6, J.D. Cowan, G. Tesauro, and J. Alspector, eds., pp.120-127(1994).
[6] M.J. Nijman and H.J. Kappen, “Symmetry Breaking and Training from Incomplete Data with Radial Basis Boltzmann Machines,” International Journal of Neural Systems, Vol.8, pp.301-316(1997).
[7] R.J. Hathaway and J.C. Bezdek, “Fuzzy c-Means Clustering of Incomplete Data,” IEEE Transactions on Systems, Man, and Cybernetics, Vol.31, No.5, pp.735-744(2001).
[8] C.P. Lim, J.H. Leong and M.M. Kuan, “A Hybrid Neural Network System for Pattern Classification Tasks with Missing Features,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol.27, No.4, pp.648-653(2005).
[9] J.A. Freeman and D.M. Skapura, Neural Networks Algorithms, Applications, and Programming Techniques, Addison-Wesley(1991).
[10]G.A. Carpenter and S. Grossberg, “Fuzzy ART: An Adaptive Resonance Algorithm for Rapid, Stable Classification of Analog Patterns,” International Joint Conference on Neural Networks, Vol.2, pp.411-416(1991).
[11]C.L. Blake and C.J. Merz, “UCI Repository of Machine Learning Databases,” Dept. of Information and Computer Science, Univ. of California, Irvine, http://www.ics.uci.edu/~mlearn/databases/(1998).
[12]S.Y. Yoon and S.Y. Lee, “Training Algorithm with Incomplete Data for Feed-Forward Neural Networks,” Neural Processing Letters, Vol.10, pp.171-179(1999).
[13]J.-S. R. Jang, C.-T. Sun and E. Mizutani, Neuro-Fuzzy AND Soft Computing, Prentice-Hall(1997).
[14]G.A. Carpenter and S. Grossberg, “ARTMAP: A Self-Organizing Neural Network Architecture for Fast Supervised Learning and Pattern Recognition,” International Joint Conference on Neural Networks, Vol.1, pp.863-868(1991).
[15]G.A. Carpenter and S. Grossberg, “Distributed ARTMAP,” International Joint Conference on Neural Networks, Vol.3, pp.1983-1987(1999).
[16]G.A. Carpenter and S. Grossberg, “Default ARTMAP,” Proceedings of The International Joint Conference on Neural Networks, Vol.2, pp.1396-1401(2003).
[17]G.A. Carpenter and S. Grossberg, “Fuzzy ARTMAP: A Neural Network Architecture for Incremental Supervised Learning of Analog Multidimensional Maps,” IEEE Transaction on Neural Networks, Vol.3, No.5, pp. 698-713(1992).
[18]G.A. Carpenter and S. Grossberg, “A Self-0rganizing Neural Network for Supervised Learning, Recognition, and Prediction,” IEEE Communications Magazine, Vol.30, No.9, pp.38-49(1992).
[19]J.C. Bezdek, J.M. Keller, R. Krishnapuram, L.I. Kuncheva and N.R. Pal, “Will the real IRIS data please stand up?,” IEEE Transactions on Fuzzy System, Vol.7, pp.368–369(1999).
[20]V.B. Rao and H.V. Rao, C++ Neural Networks & Fuzzy Logic, MIS Press(1995).