A New Approach for Fuzzy Subtractive Clustering Algorithm

The goal of clustering is to divide a given data set into distinct clusters. TheK-means and fuzzy c-means (FCM) algorithms are the most well-known clusteringtechniques. Nevertheless, there are two main issues with the K-means and FCMapproaches when it comes to actual data clustering. First, the performance of thecluster centers may be lowered due to inaccurate initial estimations. Second, it’s notalways possible to determine the number of clusters in advance. In this article, thesubtractive clustering method with a new distance measure is investigated, which doesnot have the mentioned problems. In this way, the central points of each cluster arecalculated using simple relationships, so these points will not be chosen randomly. Inorder to show the ability and efficiency of the proposed meter, a numerical example issolved at the end of the article.

Keywords:

Fuzzy clustering , Fuzzy numbers , Distance measure