Credit Risk Assessment and Prediction Algorithm of Commercial Banks

Abstract

Commercial banks are vital to the monetary system because of their indispensable part in the procedures of capital circulation (CC), capital integration (CI), capital resource allocation (CRA), then adjustment of whole societal demand and supply. Make sure you are aware of macroeconomic developments and how they could affect credit risk. Assess the loan portfolio's vulnerability to outside influences by including macroeconomic information into risk models. In this manuscript, Credit Risk Assessment and Prediction Algorithm of Commercial Banks (CRA-PCB-GRNN) is proposed. Initially the data is collected from Credit Risk Analysis Dataset. Then the collected data is fed into pre-processing utilizing Information Exchange Multi-Bernoulli Filter (IEMBF). The IEMBF is used for data normalization. Then the pre-processed data are given to Attribute-Augmented Spatiotemporal Graph Convolutional Network (AST-GCN)for predict non-performing loan rate of profitable bank. In general, AST-GCN does not express adapting optimization strategies to determine optimal parameters. Hence, the African Vulture Optimization Algorithm (AVOA)to optimize AST-GCN which accurately predict the non-functionality loan rate of profitable bank. The proposed CRA-PCB-GRNN approach is implemented in Python. The act of suggested technique examined using performance processes like Accuracy, Precision, Recall, F1-score and R2. The proposed CRA-PCB-GRNN approach contains 29.0%, 27.6%, and 26.4% higher accuracy, 25.1%, 23.4%, and 29.6% higher precision and 22.1%, 24.6%, and 20.3% higher recall related with current approaches, like, Research on Credit Risk Assessment of Commercial Banks depend on KMV method (RCRA-CB-KMV), Credit Risk Model Based on Central Bank Credit Registry Data (CRM-CBCRD-RF) also A Self-Learning BP Neural Network Calculation Procedure for Credit Risk of Commercial Bank (CRCB-BPNN) methods respectively.