Dissertation > Excellent graduate degree dissertation topics show

Research on Clustering Algorithms Based on Agglomerative Fuzzy K-Means

Author: LiXuTao
Tutor: YeYunMing
School: Harbin Institute of Technology
Course: Computer Science and Technology
Keywords: hierarchical clustering semi-supervised clustering decision tree K-nearest-neighbors Agglomerative Fuzzy K-means
CLC: TP181
Type: Master's thesis
Year: 2009
Downloads: 72
Quote: 1
Read: Download Dissertation

Abstract


With the development of information technology, human beings have accumulated more and more data. Facing such large amount of data, how to detect useful information from them is becoming one of the most important research areas of data mining, machine learning, and pattern recognition etc. K-means type clustering algorithm is one of the important tools to help to handle the problem.K-means has been widely used in different areas because of its efficiency and simplicity. However, its effectiveness has been affected by two intrinsic problems. One is that it needs to determine the number of clusters, which is also known as model selection problem. Another is that the clustering result is highly affected by the locations initial centers because of its non-convex optimization. Agglomerative Fuzzy K-means proposed by Li Mark can avoid the problems. The paper completes the following three issues based on Agglomerative Fuzzy K-means:(1) The clustering problem of data set with multilevel-density clusters (clusters with different density and hierarchical structures between them). The paper proposes a Clustering Tree algorithm based on Agglomerative Fuzzy K-means. The proposed algorithm solves the clustering problem via a tree which is induced by top-down divisively applying Agglomerative Fuzzy K-means with the combination of cluster validity and Gaussian fitness test. Experimental results have shown the effectiveness of the proposed algorithm in clustering data sets with multilevel-density clusters, meanwhile have shown that the proposed algorithm is very potential to find useful knowledge because the tree preservers the structural information between these clusters.(2) The clustering problem of data set with prior information. Based on Clustering Tree, the paper proposes a Semi-supervised Clustering Tree algorithm, which employs the prior information to clustering data background knowledge. The effectiveness of the algorithm has been proven by experimental results. (3) After applying Clustering Tree algorithm on training data set to induce a tree, a decision-tree-like classifier and a KNN-like classifier can be derived from it. The experimental results have shown that these classifiers can achieve as good accuracies as Decision Tree and traditional classification algorithms or even better.After finishing the above research issues, we propose the promising research direction in further work, which has certain significance to guide both the research and application of K-means.

Related Dissertations

  1. Fault Diagnosis Method Based on Support Vector Machine,TP18
  2. Design and Development of Teaching Quality Assessment System Based on Data Mining,TP311.13
  3. An Improvement of Cluster on Phylogenetic Profiling Method,TP311.13
  4. Based on the User’s Perceptual Needs’ Handle GPS Design,TN967.1
  5. Web-based social network search method to judge the identity of names,TP393.09
  6. Study on the Decision Tree Classification Algorithm and Its Application Based on Rough Set Theory,TP18
  7. Decision Tree Based on Generalized Information Entropy Model and Its Performance Evaluation,TP18
  8. P2P Traffic Identification Method,TP393.06
  9. Study on the Decision-Making Model of the Hepatitis B Immunization among the People Over 15 Years Old in China,R392.1
  10. Study on the Economic Burden of Hepatitis B-related Diseases in Guangzhou and the Strategies of Hepatitis B Vaccination for Population Aged 1～14 in China,R186
  11. Research on the Application of Data Mining Technology in the Regulating Matriculation for Postgraduate,TP311.13
  12. Research of Web Text Classification Based on Decision Tree Classification Algorithm,TP391.1
  13. Research and Implementation of Decision Tree-based Short-term Load Forecasting System,TM715
  14. Research and Application on Voyage Knowledge Discovery,F552.6
  15. Loading Optimization Method Based on Two Level Clustering,U492.22
  16. A certain type of active radar -guided missiles Universal Automatic Test Equipment Research,TJ760.6
  17. Study on Extraction of Coniferous Forest Information in Southern China,TP79
  18. Chinese hepatitis B vaccine effectiveness tracking Decision,R186
  19. Multi-level Selective Ensemble Learning Based on Multi-task Learning,TP181
  20. The Application of SVM to Decision Tree Induction,TP18
  21. Fuzzy Rough Set Attribute Reduction and Multi-fuzzy Decision Information Fusion,TP18

CLC: > Industrial Technology > Automation technology,computer technology > Automated basic theory > Artificial intelligence theory > Automated reasoning,machine learning
© 2012 www.DissertationTopic.Net  Mobile