Achieving Balance in Clusters- A Survey
Divya Saini, Manoj K. Singh · 2015
K-means Clustering Algorithm, among the various clustering algorithms proposed till date, has proved its superiority by its simplicity and usability. However, it is prone to a number of limitations, one being lack of balance in clusters. Clusters obtained, if balanced, will result in equally sized clusters thereby distributing the load equally and enhancing the quality of clustering. The existing clustering algorithms aiming at optimizing the traditional k-means are limited to working on the bad initialization problem or the local optimum problem of k-means. This paper discusses some of the relevant research works in the direction of obtaining balance in clusters, its need and approaches.