Fusion of Kohonen Maps Ranked by Cluster Validity Indexes
In this study, a new approach to Kohonen Self-Organizing Maps fusion is presented: the use of modified cluster validity indexes as a criterion for merging Kohonen Maps. Computational simulations were performed with traditional dataset from the UCI Machine Learning Repository, with variations in map size, number of subsets to be merged and the percentage of dataset bagging. The fusion results were compared with a regular single Kohonen Map. In some selected parameters, the proposed method achieves a better accuracy measure.
KeywordsFusion Self Organizing Maps Validity Index
Unable to display preview. Download preview PDF.
- 3.Perrone, M.P., Cooper, L.N.: When networks disagree: ensemble methods for hybrid neural networks. In: Neural Networks for Speech and Image Processing, pp. 126–142. Chapman and Hall (1993)Google Scholar
- 7.Gonçalves, M.L., De Andrade Netto, M.L., Costa, J.A.F., Zullo, J.: Data clustering using self-organizing maps segmented by mathematic morphology and simplified cluster validity indexes: an application in remotely sensed images. In: International Joint Conference on Neural Networks, IJCNN 2006, pp. 4421–4428 (2006)Google Scholar
- 8.Georgakis, A., Li, H., Gordan, M.: An ensemble of SOM networks for document organization and retrieval. In: International Conference on Adaptive Knowledge Representation and Reasoning (2005)Google Scholar
- 14.DeLooze, L.L.: Attack characterization and intrusion detection using an ensemble of self-organizing maps. In: 2006 IEEE Information Assurance Workshop, pp. 108–115 (2006)Google Scholar
- 22.Bache, K., Lichman, M.: Machine Learning Repository. University of California, Irvine, School of Information and Computer Sciences (2013), http://archive.ics.uci.edu/ml