Staged-adaptive data clustering in fuzzy min-max neural network

In this paper, a data clustering approach called staged-adaptive data clustering in fuzzy min-max neural net-work(SFMM) is proposed. In SFMM, a staged-adaptive process is designed to improve the performance of the he fuzzy minmax neural network. With the added staged-adaptive process, the accuracy and stability are improved. Meanwhile, the process of the hyperbox expansion has been modified. With the modification of the fuzzy min-max neural network, the number of the hyperbox can be controlled in some degree. For evaluation purposes, several data sets from UCI have been utilized. At the same time, some related clustering algorithms have been added to make a comparison with SFMM. The experimental results shows that SFMM has better performance.