An Approach to Determine the Number of Clusters for Clustering Algorithms

详细信息    查看全文

• 作者：Dinh Thuan Nguyen (22)
  Huan Doan (23)
  
• 关键词：clustering ; selecting the number of clusters ; fuzzy c ; means
• 刊名：Lecture Notes in Computer Science
• 出版年：2012
• 出版时间：2012
• 年：2012
• 卷：7653
• 期：1
• 页码：495-504
• 全文大小：342KB
• 参考文献：1. Yan, M.S., Wu, K.L., Yu, J.: A novel fuzzy clustering algorithm. In: IEEE In’t Symp. on Computational Intelligence in Robotics and Automation, pp. 647-52 (2003)
  2. Pham, D.T., Dimov, S.S., Nguyen, C.D.: Selection of K in K-means clustering. In: Proc. IMechE, Mechanical Engineering Science, vol.?219, pp. 103-19 (2005)
  3. Yin, Z., Tang, Y., Sun, F., Sun, Z.: Fuzzy Clustering with Novel Separable Criterion Tsinghua Science and Technology, pp. 50-3 (2006) ISSN 1007-0214 09/2011
  4. Haizhou, W., Mingzhou, S.: Optimal k-means Clustering in One Dimension by Dynamic Programming. The R Journal?3(2), 29-3 (2011) ISSN 2073-4859
  5. Mohammad, F.E., Wesam, M.A.: Initializing K-Means Clustering Algorithm using Statistical Information. In’t Journal of Computer Applications, 51-5 (2011)
  6. Santhi, M.V.B.T., Sai Leela, V.R.N., Anitha, P.U., Nagamalleswari, D.: Enhancing K-Means Clustering Algorithm. International Journal of Computer Science & Technology, IJCST?2(4), 73-7 (2011) ISSN: 2229-4333
  
• 作者单位：Dinh Thuan Nguyen (22)
  Huan Doan (23)
  
  22. University of Information Technology, VNU-HCM, Vietnam
  23. EnterSoft Software Solution Joint Stock Company, HCM, Vietnam
  
• ISSN：1611-3349

文摘

When clustering a dataset, the right number k of clusters is not often obvious. And choosing k automatically is a complex problem. This paper first reviews existing methods for selecting the number of clusters for the algorithm. Then, an improved algorithm is presented for learning k while clustering. The algorithm is based on coefficients α, β that affect this selection. Meanwhile, a new measure is suggested to confirm the member of clusters. Finally, we evaluate the computational complexity of the algorithm, apply to real datasets and results show its efficiency.