Composition of probabilistic preferences with an empirical approach in multi-criteria problems

Abstract This paper aimed at presenting an Empirical Approach to the Composition of Probabilistic Preferences (CPP) method. The CPP method is used to choose, sort and classify alternatives in multi-criteria problems. The CPP has explored a wide variety of applications from continuous probability distributions. However, the use of empirical probabilities, without the need to know or assume a joint probability function that gives rise to preferences, can help to deal with certain types of problems with ordinal preference scales. The empirical approach is applied here in three cases, with the purpose of illustrating its main characteristics and limitations. The proposal proved to be satisfactory for treating appraisers' preferences, especially when the number of alternatives is reduced in relation to the options of the scale of assessment points and the number of appraisers, as they contribute to increase the discriminatory capacity of the results.