Hierarchical fuzzy rule-based model for managing credit rating of a banking ecosystem client

The study proposes a hierarchical fuzzy rule-based model that allows one to determine the assessment of the state of a client of the banking ecosystem and calculate his/her credit rating. Additionally, the proposed model allows one to form multiple intermediate assessments. These intermediate assessments are obtained at the output of individual fuzzy rule-based models that form a hierarchical structure. The use of intermediate assessments allows one to determine the groups of parameters that influenced the value of the aggregated assessment and the value of the client’s credit rating. If the proposed model produces low values of the credit rating, the analysis of intermediate variables begins to identify the causes. As a result of the analysis, a set of input variables is formed, with the help of which the reasons for assigning a certain credit rating are explained. To correct the low value of the credit rating, control actions are formed. A feature of such actions is the involvement of the client in the business processes of the banking ecosystem. The experiment confirmed the possibility of the proposed theoretical provisions to distribute the objects of analysis by state classes and determine the values of aggregated assessments for them depending on various combinations of input parameter values. The possibility of using the proposed model to explain the obtained results by forming maps of the client’s state and predicting the result of applying control actions on the client’s state has been confirmed.