The K-means algorithm is usually widely used in cluster analysis, but it is easily disturbed when dealing with data containing outliers.