基于机器学习方法的银行贷款决策模型  

Bank Loan Decision Model Based on Machine Learning

作　　者：王辉 杨航 刘琦 WANG Hui;YANG Hang;LIU Qi(Yellow River Communications College,Jiaozuo 454950,China)

机构地区：[1]黄河交通学院,河南焦作454950

出　　处：《电脑与电信》2025年第1期70-75,共6页Computer & Telecommunication

摘　　要：随着网络借贷的迅速发展以及市场的饱和,银行贷款业务的客户数量越来越少,甚至让银行陷入获取贷款客户难的窘境。因此,银行客户的贷款分析与预测至关重要。探讨了银行贷款客户与非贷款客户的差异,采用描述性统计分析,并通过斯皮尔曼单因素分析和随机森林算法进行特征选择,以分析用户特征之间及其与贷款意愿的相关性。最终,利用随机森林、逻辑回归、朴素贝叶斯、梯度提升树和XGBoost等算法构建多套贷款预测模型,通过效果对比确定了最佳模型。研究结果表明,所建立的模型具备较高的准确性,能有效支持银行决策者预测客户贷款意愿。With the rapid development of online lending and the saturation of the market,the number of customers in the bank loan business is getting less and less,and even makes it difficult for the bank to obtain loan customers.Therefore,the loan analysis and prediction of bank customers are crucial.This paper explores the differences between bank loan customers and non-loan customers,using descriptive statistical analysis and feature selection through Spearman univariate analysis and random forest algorithm to ana‐lyze the correlation between user characteristics and their willingness to lend.Finally,random forest,logistic regression,naive Bayes,gradient lifting tree and XGBoost algorithms are used to build multiple loan prediction models,and the best model is deter‐mined by effect comparison.The results show that the model has high accuracy and can effectively support bank decision makers to predict customers'loan willingness.

关 键 词：机器学习模型 贷款预测 模型可视化 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]