| The purpose of this article was to introduce the basic concepts, calculation methods, two examples and their SAS implementations for the cluster analysis of disordered samples. The basic concepts included the ideal scheme or optimal scheme, negative ideal scheme or worst scheme, co-trending, Euclidean distance and comprehensive index. The calculation method involved the calculation principle of Topsis method, the implementation steps and formulas of Topsis method. The data in the two examples were respectively "expert evaluation results of 10 comprehensive evaluation indexes reflecting the work quality of a hospital from 1995 to 2004" and "investigation results of 4 indexes of physical development of urban male students aged 18~25 in 16 provinces and cities of China in a certain year". Using SAS software, the data in the two examples were analyzed by cluster analysis of disordered samples, and then the results of classification were given, and the output results of SAS were explained. |