Adaptive fuzzy rule-based systems for credit rating analysis

The paper presents an analysis of credit rating process by using adaptive fuzzy rule-based systems. First, previous studies in credit rating analysis are reviewed. The disadvantage of the models used in these studies consists in the fact that it is difficult to extract understandable knowledge from them. This problem appears to be crucial also because the use of natural language is typical for the credit rating process. It can be solved by using fuzzy logic, enabling its user to model the meaning of natural language words. Therefore, the model based on the use of fuzzy logic is designed to classify US companies and municipalities into the credit rating classes obtained from notable rating agencies. The model includes data preprocessing, the selection process of input variables, and the design of various adaptive fuzzy rule-based systems for classification. The selection of input variables is realized using genetic algorithms. The objective of this process is to select only significant variables in order to improve the performance of adaptive fuzzy rule-based systems. Input variables are extracted from financial statements and capital markets in line with previous studies. These variables represent the inputs of adaptive fuzzy rule-based systems, while the credit rating classes stand for the outputs. The results show that the credit rating classes assigned to bond issuers can be classified with a high accuracy rate using a limited subset of input variables. Moreover, the use of fuzzy logic makes it possible to interpret the model properly. Thus, credit rating process can be replicated in an understandable way.