A Simplified Interval Type-2 Fuzzy Implementation for Financial Credit Decision

Abstract

Credit assessment of potential customers with the help of their previous history of credit transaction is one of the main issues in financial credit approval system. The credit rating of the customer shows that the financial stability of the individual or firm. Based on financial stability, the bankers can approve their credit grant. The basic factors that affect the credit rating of the customer is history of payment, the unsettled amount, period of credit history, types of credit and many other factors. The creditworthiness of the customer is assessing based on result obtained from these factors. The prime objective of the credit approval system is to avoid loss of amount that may be associated with an incorrect decision. To avoid such type of decisions, it requires designing of credit rating models for credit and their risk analysis. This type of models benefited the bankers to approve the credit grant or not. The bank credit system is a binary classification system that classifies the customer either the good or bad based on their previous credit history. In this context, several fuzzy classification systems have been designed to classify the customer. In this article, we have designed a simplified interval type-2 fuzzy system for financial credit decision using two different membership function based approaches and compared the performance in terms of accuracy of classification.