138 A Novel Approach towards Bisecting K-means Clustering Algorithm Parallelism

Considering the insufficiency of clustering speed which exists in the selecting the initial centroids of bisecting K-means (BKM) clustering algorithm, the idea of selecting the two patterns with distance maximum as the initial cluster centroids is implemented. The results show that the clustering speed is better than that of BKM. An in-depth study and analysis is carried out on how to accelerate clustering in parallel computing system. According to the characteristics of BKM, the parallelism algorithm based on data parallelism is put forward. Undertaken experimental results show that the improvement of algorithm get ideal speed up performance and efficiency.