An improved K-means algorithm with meliorated initial center

K-means algorithm is commonly used in clustering algorithms to find clusters due to its simplicity of implementation and fast execution. However, classical K-means algorithm is sensitive to the initial clustering center. To improve the performance of K-means algorithm, a new method for initial center selection is presented in the paper. The method first finds the largest cluster, then makes the cluster to split by two data objects which have the maximum distance as the first clustering centers and does the above steps repeatedly until the specified number of clustering centers is obtained. Compared to the original algorithm, the experiment results on KDD CUP99 dataset show that the improved algorithm has a better clustering effect.