| The purpose of this article was to introduce the basic concepts, computational methods, two examples and their SAS implementations related to cluster analysis of disordered samples. The basic concepts included rank transformation, high-benefit indicators or benefit-type indicators, expected target indicators, slightly higher optimal indicators, and concepts related to the rank-sum ratio method. The computational methods involved formulas corresponding to two approaches. The first was when the denominator was the total number of data points, under the assumption that the evaluation indicators had no weight coefficient or had weight coefficients, the formulas for calculating the rank-sum ratio of each evaluation object were presented. The second was when the denominator was the sum of all ranks, under the assumption that the evaluation indicators had no weight coefficient or had weight coefficients, the formulas for calculating the rank-sum ratio of each evaluation object were also presented. The data in the two examples were "5 economic indicators and their values in different density forms" and "a sample set of agricultural productivity rating indicators", respectively. Using SAS software, a comprehensive evaluation of the evaluation objects in the two examples was conducted, including the results of ranking and grading the evaluation objects. |