An efficient algorithm for evaluating the cluster analysis

It is a main task of exploratory data mining, and a common technique for statistical data analysis, used in many fields, including machine learning, pattern recognition, image analysis, information retrieval, and bioinformatics. This paper proposes an efficient algorithm used to evaluate the results of cluster analysis in classifying the patients with the tumor. The Prediction Strength, combined with the conceptions of the median and the 95% reference ranges, is used to compare the classified patients categorized by various distances between clusters. The method, to a certain extent, fills up the deficiency of the cluster analysis, and has shown a successful application to classifying the patients with leukemia.