A Performance Neighborhood Distance (ndist) Between K -Means and SOM Algorithms

Abstract

Clustering is an important means of data mining based on separating data categories by similar features. This paper aims to compare the performance of neighborhood distance (ndist) between K-Means and Self-Organizing Maps (SOM) algorithms. The sample in this study is rainfall datasets, from 13 Stations in East Kalimantan. The performance of ndist was used Euclidean Distance. This paper outlines and presents the comparison performance of ndist of K-Means and SOM for analyzing and clustering rainfall datasets. The performances of these algorithms are compared based on the ndist values. The findings of this study indicated that the K-Means has been proved to be effective in ndist by using centroid concept better than SOM algorithm. This paper is concluded by recommending some future works that can be applied in order to improve the ndist of K-Means and SOM.