Unsuperised classification by soft computing techniques: algorithms of fuzzy -means clustering

We overview methods of fuzzy -means clustering as a representative techniques of unsupervised classification by soft computing. The basic framework is the alternate optimization algorithm originally proposed by Dunn and Bezdek is reviewed and two more objective functions are introduced. An additional variable of controlling volume size is included as an extension. Moreover a method of the kernel trick for obtaining nonlinear cluster boundaries is moreover considered and a simple numerical example is shown.