An Automatic Multi-Objective Clustering Based on Hierarchical Method
Just optimizing a single objective function or need to know the exact number of clusters in advance is the choice of most clustering methods. However, less knowledge of the data set to be clustered makes it difficult to select the appropriate number of clusters. Motivated by this, we propose an automatic multi-objective clustering based on hierarchical method (AMOH-Cluster), which can not only automatically calculate the optimal number of clusters but also divide all data sets properly based on intra-cluster data compactness and inter-cluster data connectivity. The proposed algorithm has advantages of providing higher clustering accuracy and requiring only a few parameters. As shown in the experiment, the comparison with the known multi-objective clustering algorithms proves that the proposed algorithm provides a solution with higher accuracy and optimal clustering number in various clusters of artificial data sets.
KeywordsMulti-objective clustering Automatic clustering Hierarchical method
This work was supported by the Natural Science Foundation of China (No. 61472231). Natural Science Foundation of China (No. 61502283). Natural Science Foundation of China (No. 61640201).
- 4.Dempster, A. P., Laird, N. M., & Rubin, D. B. (1977). Maximum likelihood from incomplete data via the EM algorithm. Journal of the Royal Statistical Society: Series B: Methodological, 39(1), 1–38.Google Scholar
- 5.Handl, J., & Knowles, J. (2004). Evolutionary multiobjective clustering. International conference on parallel problem solving from nature. Springer Berlin Heidelberg, London (pp. 1081–1091).Google Scholar
- 11.Chen, Y., Tang, C., Zhu, J., et al. (2007) Clustering without prior knowledge based on gene expression programming 2007. ICNC 2007. Third International Conference on Natural Computation, IEEE, 3, 451–455Google Scholar
- 12.Bandyopadhyay, S., & Pal, S. K. (2007). Classification and learning using genetic algorithms: applications in bioinformatics and web intelligence (Vol. 31). Springer Science & Business Media.Google Scholar