Predicting credit risk using discriminant analysis

Abstract

The study aims to determine the effectiveness of the discriminant analysis method to make a single standard that determines the capacity of the institution requesting the loan to repay or not, as controlling the risk is a priority for banking institutions in view of the volume of loans granted, which prompts them to include a mechanism to control it within the auditing methods and Management, the increase or decrease in the volume of loans granted that appears in bank statements is the focus of attention, as it affects the bank’s ability to maintain a certain level of liquidity in order to preserve its solvency. The banking system relies on financial analysis and the requirement of guarantees as two classic tools to reduce loan risk. However, it tends to use applied explanatory statistics methods to measure loan risk, which are methods that exploit financial analysis methods represented by financial ratios to obtain a single indicator to measure loan risk. The results of using discriminant analysis and Bayesian classifier showed a correct classification rate of up to 97%, which is embodied in 45 out of the total 46 institutions of the sample