An Improved Approach for k-Means Using Parallel Processing

Serial execution of K-means algorithm on large dataset takes more execution time and does not give accurate results. Parallel processing is one of the ways to improve the performance of K-Means algorithm. But the execution time and accuracy is largely dependent on selection of initial cluster centers. In this paper, parallel processing of K-Means is proposed using an initialization method to originate initial cluster centers, which not only reduces the execution time but also gives accurate results.