Improved K-mediods algorithm

The scientific data mining becomes one of the hot topics in the research and application of the data base system.The clustering is an important part of the data mining.The result of the K-mediods is affected by the initial center points. And it always terminates in the local optimization.If the initial center points are dense and these points are dissimilar each other the time of the algorithm execution is reduced and the quality of the results is improved.