Study of efficiency k-means clustering using Z-test proprieties

摘要

In data mining, clustering is a technique of regrouping similar objects with common proprieties in some clusters. K-means algorithm is the basic of clustering technique; it is the most widely used algorithm for diverse applications. This paper studies and analyses the efficiency of extending k-means results of a perfect sample set, to different sets by using Z-test proprieties, this is based on the K-means algorithm and working with the uniform distribution of data points. The objective is to achieve better clustering with reduced complexity, by keeping the same number of clusters and maximized belonging points to each cluster. The result of the proposed clustering method was investigated during different calculus of Z value for different input data points.