- 著者
-
Miyamoto Sadaaki
Suizu Daisuke
- 雑誌
- Proceedings of the 1st International Conference on Fuzzy Systems and Knowledge Discovery
- 巻号頁・発行日
- vol.FSDK'02; 2, pp.656-660, 2002-11
Algorithms of fuzzy c-means clustering with kernels employed\nin nonlinear transformations into high dimensional\nspaces in the support vector machines are studied. The objective\nfunctions in the standard method and the entropy\nbased method are considered and iterative solutions in the\nalternate optimization algorithm are derived. Explicit cluster\ncenters in the data space are not obtained by this method\nin general but fuzzy classification functions are useful which\nhave much more information than crisp clusters in the hard\nc-means. Numerical examples using radial basis kernel functions\nare given.