Efficient Determination of Clusters in K-Mean Algorithm Using Neighborhood Distance

Abstract: Clustering is a well known data mining technique which is used to group together data items based on similarity property. Partitional clustering algorithms obtain a single partition of the data instead of a clustering structure. K-mean clustering is a common approach; however one of its drawbacks is the selection of initial centroid points randomly because of which algorithm has to re-iterate number of times. This paper first reviews existing methods for selecting the number of clusters as well as selecting initial centroid points, followed by a proposed method for selecting the initial centroid points and the modified K-mean algorithm which will reduce the number of iterations and improves the elapsed time