跳到主要內容

臺灣博碩士論文加值系統

(3.236.84.188) 您好!臺灣時間:2021/08/05 01:22
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:蕭介銘
研究生(外文):Chieh-ming Hsiao
論文名稱:以自適應共振網路與概念階層為基礎之非監督式分群演算法
論文名稱(外文):Unsupervised Clustering Algorithm based on Concept Hierarchy and Adaptive Resonance Theory Neural Network
指導教授:許中川許中川引用關係
指導教授(外文):Chung-Chian Hsu
學位類別:碩士
校院名稱:國立雲林科技大學
系所名稱:資訊管理系碩士班
學門:電算機學門
學類:電算機一般學類
論文種類:學術論文
論文出版年:2004
畢業學年度:92
語文別:中文
論文頁數:61
中文關鍵詞:資料探勘分群演算法種類型資料概念階層自適應共振網路
外文關鍵詞:K-prototypecategorical dataART2Data MiningConcept hierarchy.
相關次數:
  • 被引用被引用:1
  • 點閱點閱:131
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
長久以來,群聚分析一直是一項相當重要的資料分析工具,截至目前為止已有許多的分群演算法被提出,這些演算法對於連續型屬性資料分群大都有良好的結果,但對於混合型資料分群的結果卻不盡理想,其主要原因在於目前大部分的分群演算法,無法合理計算種類型資料的相似度,因為它比單獨針對數值型屬性或種類型屬性作運算更為複雜,使得這些演算法無法被廣泛應用在資料探勘領域中。有鑑於此,本研究藉由概念階層樹表達種類型資料的相似度,來改善傳統的分群演算法無法合理計算種類型資料彼此相似度的缺點,將其整合在自適應共振網路理論裡,使其能夠同時對種類型和數值型混合資料作分群,以致於能夠更廣泛的被應用在各個領域,在此我們將此演算法稱為M-ART。實驗部份,我們藉由人工方式產生三個資料集,以及UCI資訊及電腦科學系所收集之成人收入資料集,進行一系列實驗,將結果與傳統的自適應共振網路,和能夠分群混合型資料的k-prototype演算法做比較,結果證明,M-ART比ART2及K-prototype演算法能更合理且正確的將混合型資料作分群,且其執行時間是與資料量呈線性比例增加。
Cluster analysis has been an important tool for data analysis all the time and there are many clustering algorithms being published so far. Generally, these algorithms can get good clustering results, especially on numerical data. However, they didn’t work well on mixed data. The reason is that most of clustering algorithms so far couldn’t reasonably compute the similarity of categorical data owing to no proper representation scheme. Consequently, a traditional clustering algorithm needs to be extended before being widely applied to mining various types of data. This study aims to study the representation of the similarity of categorical data via concept hierarchy trees and its application, in particular, to improve the conventional ART algorithm for handling hybrid data. The revived algorithm is referred to as M-ART. The experimental outcomes show that M-ART can cluster mixed data sets more reasonably and precisely than ART2 and K-prototype. The compute time is liner
摘要 I
ABSTRACT II
誌謝 III
目 錄 IV
圖 目 錄 VI
表 目 錄 VII
表 目 錄 VII
第一章 緒論 9
1.1 研究動機 9
1.2 研究目的 10
1.3 研究範圍及限制 10
1.4 研究步驟 10
第二章 文獻探討 12
2.1 群聚分析 12
2.2 分群演算法應具備之條件 12
2.3 自適應共振理論網路 13
2.3.1 自適應共振理論網路演算法 14
2.3.2 自適應共振網路計算種類型資料的缺點 16
2.3.3 自適應共振理論網路相關應用 16
2.4 種類型資料間距離計算的相關研究 17
2.5 分群品質的評估 21
第三章 研究架構 22
3.1 概念階層表達種類型資料 22
3.2 改良式自適應共振網路 23
3.3 網路架構 24
第四章 實驗結果 29
4.1 資料前置處理 29
4.2 人工資料集實驗 29
4.2.1 人工資料集實驗1 30
4.2.2 人工資料集實驗2 32
4.2.3 人工資料集實驗3 36
4.3 UCI真實資料集實驗 38
4.3.1 實驗步驟 39
4.3.2 亂度分析 39
4.3.3 種類型屬性CU值分析 41
4.3.4 數值型屬性分析 42
4.3.5 M-ART群聚原型的表達 43
4.3.6 K-Prototype群聚原型的表達 45
4.3.7 K-Prototype與M-ART群聚原型之比較 47
4.3.8 執行時間 48
4.3.9 穩定度實驗 48
第五章 結論與未來研究 50
5.1 結論 50
5.2 未來研究 50
第六章 參考文獻 52
附錄一 M-ART對各資料集分群結果的原型 55
附錄二 K-Prototype對各資料集分群結果的原型 58
1.蘇木春、張孝德,機器學習:類神經網路、模糊系統以及基因演算法則,全華科技圖書股份有限公司,2002。
2.Barbara, D., J. Couto and Y. Li, COOLCAT: an entropy-based algorithm for categorical clustering, Proceedings of the eleventh international conference on Information and knowledge management, pp. 582-589, 2002.
3.Carpenter, G. A. and S. Grossberg, ART 2:Self-organization of stable category recognition codes for analog input patterns, Appl. Opt. Vol. 26, pp. 4919-4930, 1987
4.Carpenter, G.. A., Carpenter, G.. A., and S. Grossberg, Grossberg, S., The ART of Adaptive Pattern Recognition by a Self-Organizing Neural Network, Computer Vol., 21, No. 3, pp. 77-88, 1988.
5.Cheung, D.W., H.Y. Hwang, A.W. Fu, and J. Han, Efficient Rule-Based Attribute-Oriented Induction, Journal of Intelligent Information Systems, Vol. 15, pp. 175-200, 2000.
6.Distante, C., S. Pietro, and V. Lorenzo, “Odor discrimination using adaptive resonance theory”, Sensors and Actuators, Vol. 69, No. 3, pp. 248-252, 2000.
7.Guda, S., R. Rastogi, K. Shim, ROCK: A Robust Clustering Algorithm for Categorical Attributes, Proceedings of the IEEE Conference on Data Engineering, 1999.
8.Han, J. and M. Kamber, 2001, “Data Mining Concepts and Techniques,” San Francisco: Morgan Kaufmann.
9.Han, J. and Y. Fu, Dynamic Generation and Refinement of Concept Hierarchies for Knowledge Discovery in Databases, Workshop Knowledge Discovery in Databases, pp. 157-168, 1994.
10.Hsien, C.C., K.L. Pheng, and Y. Wei., “A strategy for acquiring customer requirement patterns using laddering technique and ART2 neural network,” Advanced Engineering Informatics, Vol. 16, Issue: 3, pp. 229-240, July, 2002.
11.Hsu, Chung-Chian, Generalizing self-organizing maps for categorical data, submitted for publication, 2003.
12.Huang, Z., A Fast Clustering Algorithm to Cluster Very Large Categorical Data Sets in Data Mining, SIGMOD’97 Data Mining Workshop, 1997.
13.Huang, Z., Clustering Large Data Sets With Mixed Numeric and Categorical Values, In Proceeding of The First Pacific-Asia Conference on Knowledge Discovery and Data Mining, 1997.
14.Kohonen, T., The Self-Organizing Map, Proceeding of the IEEE, pp.1464-1480, 1990.
15.Liao, T.W., and K.S. Lee, “Integration of a Feature-Based CAD System and an ART1 Neural Model for GT Coding and Part Family Forming”, Computers & Industrial Engineering, Vol. 26, No. 1, pp. 93-104, 1994.
16.Massey, L., “On the quality of ART1 text clustering”, Neural Networks, Vol. 16, No. 5, pp. 771-778, 2003.
17.Mulder, S. and I.I. Wunsch, C. Donald, “Million city traveling salesman problem solution by divide and conquer clustering with adaptive resonance neural networks”, Neural Networks, Vol. 16, pp. 827-832, 2003.
18.Murphy, P. M. and D. W. Aha., UCI repository of machine learning database, http://www.ics.uci.edu/~mlearn/MLRepository.html, 1992.
19.Waikeung, F. and L. Yunhui, “Adaptive categorization of ART networks in robot behavior learning using game-theoretic formulation”, Neural Networks, Vol. 16, No. 10, pp. 1403-1420, 2003.
20.Xiang-Rong Yang, Jun-Yi Shen and Qiang Liu, A Novel Clustering Algorithm Based on Weighted Support and Its Application, Proceeding of the First International Conference on Machine Learning and Cybemetics, 2002.
21.Yun, Ching-Huang, Kun-Ta Chuang, and Ming-Syan Chen, Using Category-Based Adherence to Cluster Market-Basket Data, IEEE International Conference on Data Mining, pp. 546-553, 2002.
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top