Font Size: a A A

Research On Risk Evaluation Of Bank Auto Finance Loan Based On Random Forest Model

Posted on:2020-12-23Degree:MasterType:Thesis
Country:ChinaCandidate:Y P DengFull Text:PDF
GTID:2492306113464634Subject:Finance
Abstract/Summary:PDF Full Text Request
With the continuous development of the bank’s automobile consumer financial loan business,the market competition has become increasingly fierce,and the non-performing rate of bank auto finance loans has risen rapidly.Therefore,optimizing the risk assessment method for auto finance loan customers and reducing the non-performing rate of customers’ auto finance loans are urgent needs of banks.This paper studies the risk assessment of bank auto finance loans based on the problems encountered by banks in the risk assessment of auto finance consumer loans.Through sufficient literature research and theoretical analysis,this paper innovatively introduces third-party customer information and proposes a new credit evaluation index system for the shortcomings of the current customer credit evaluation index system,and finally builds a PSO-based optimization algorithm.A random forest classification model to achieve an accurate assessment of the risk of auto financing loans.This paper uses the real data of Bank A to demonstrate the good performance of the new indicator system and model.The specific work of this paper is as follows:(1)Analyze the problem.This paper takes Bank A as the research object,analyzes the development,problems and causes of A bank’s current auto loan business;(2)builds an indicator system.This paper designs a new evaluation of automobile consumer loans,including basic information,loan information,financial transactions,and Union Pay third-party information,and other 29 indicators;(3)model construction.This paper proposes a random forest classification model based on PSO algorithm to predict the default risk of automobile loans;(4)empirical research.Based on the real data of Bank A,this paper verifies that the random forest model based on PSO algorithm is better than the logistic regression model,and compares the importance of the variables output by random forest model to verify the indicators designed in this paper.The effectiveness of the system.The research of this paper has important practical significance for A Bank to improve the risk evaluation system of auto finance loans,and it also provides reference for other similar banks to improve their auto loan credit risk evaluation system.
Keywords/Search Tags:auto finance loan, credit risk assessment, random forest, particle swarm optimization
PDF Full Text Request
Related items