The combination of rival penalized competitive learning and Self-Organizing Map in a class of data clustering

Either rival penalized competitive learning (RPCL) algorithm or self-organizing map (SOM) algorithm can automatically perform data clustering without knowing the exact number of clusters. However, the performance of RPCL is sensitive to the pre-selection of the rival de-learning rate, and SOM algorithm spends lots of time to learn because of many factors when it is applied in data clustering problem. In this paper, we attempt to combine the two algorithms by various methods in which we discuss the problem occur and offer the simulation result in details. The purpose of this paper is to try a series of methods to modify and debug the combination of this two algorithms such that an appropriate algorithm is given later for some special data clustering such as the experiment in this paper.