Comparison of Result Clustering Study Case Posyandu With The Scalable K Means ++ Clustering Method

Abstract

Application of data grouping aims to group data unsupervised, in this study comparing the results of the grouping with the K mean clustering method, K Means ++ clustering method and the Scalable K Means ++ clustering method. Based on the test results by analyzing the iteration error value, the results of the analysis show that the K Means ++ clustering and Scalable K Means ++ clustering method will produce less error values when compared to the K Means Clustering method. The data used as the basis of analysis in this study is based on data from Posyandu Rajawali Singosari in Malang. The initial initialization value of the centroid can be determined or randomly and is very influential for the data grouping process. Calculation analysis program used scilab programming and the error results with the graph of the minimum value. Result in test data, error value test data 1 get Scalable K Means ++ clustering error minimum 0,07, test data 2 get error value minimum K Means ++ Clustering 0,15, test data 3 get error value minimum 0,005 at metode Scalable K Means Clustering, test data 4 get error value minimum 0,15 at K Means ++ Clustering.