Rough set theory in the classification of loan applications

The article presents the results of the study, which is a fragment of the work carried out under the project entitled “Hybrid system for intelligent diagnostics of prognostic models”, co-financed through the National Centre for Research and Development from the European Regional Development Fund. The aim of the study was to analyse the effectiveness of methods based on rough set theory in binary classification tasks, on the example of cash loans granted. In accordance with the assumptions for the construction of the hybrid system, it was assumed that an important criterion for the selection of methods included in various scenarios of implemented research procedures is their automation, starting from machine selection of features, through machine discretization and model construction until obtaining classification results. Based on the results of preliminary tests, the research procedure was established, which included the use of: 1) the method of random undersampling in order to balance training data set, 2) heuristic MD algorithm (Maximum Discernibility) for discretization and reduction of variables, and 3) LEM 2 algorithm for generating the minimal set of rules for decision-making. In subsequent scenarios, the basic research procedure was extended by the stage of initial reduction of features using two filtration methods, CFS and SU-FCBF, and using the results of the expert approach commonly used in the construction of logistic regression models. By adding the number of sub-scenarios, a total of 242 different classification results were obtained. These results showed that the use of filtration methods resulted in a significant improvement in the results of rule classification.