An Optimization Algorithm of Selecting Initial Clustering Center in K - means

Abstract

Abstract:The traditional stand-alone K-means clustering algorithm has the limitation of time consumption and memory overflow when dealing with large-scale data. Although this problem is solved with the help of MapReduce framework. However, the clustering accuracy effect is not stable due to the selection of initial clustering center. Therefore, this paper presents an algorithm for optimizing the initial clustering center in K-means by using several equal-scale sampling, calculating the local density and selecting the optimal initial clustering center. The experimental results show that the optimized algorithm shortens the clustering time and improves the accuracy and stability of clustering procedure in K-means.