基于稀疏逻辑回归的信用风险评估模型

作     者：王丽华 彭定涛 

作者机构：贵州大学数学与统计学院贵州 贵阳 

出 版 物：《电子商务评论》 (E-Commerce Letters)

年 卷 期：2025年第14卷第1期

页      面：1354-1360页

学科分类：12[管理学] 02[经济学] 0202[经济学-应用经济学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 020204[经济学-金融学（含∶保险学）] 

基　　金：国家自然科学基金(12261020) 贵州省科技计划(黔科合基础ZK一般009)和贵州省高层次留学人才创新创业择优资助重点项目(03) 

主　　题：信用风险 稀疏优化 逻辑回归 

摘      要：随着经济的持续增长和金融科技的不断发展,个人信贷作为一种满足消费需求的金融工具,其市场规模自然随之扩大。受到经济下行压力、不良贷款行为增加与各种突发变故的影响,个人信贷违约率逐渐上升,一个完善且高效的个人信用评估模型其重要性不言而喻。在信用评估过程中,通过一系列的具体指标和因素去判断个人的信用风险,在庞大的市场规模下,需要巨量的资源投入。本文提出了一种基于稀疏优化的逻辑回归模型,其能在保持一定准确度的情况下快速地得出个人风险评估结果。最后通过真实数据,验证所提出稀疏逻辑回归模型的有效性。With the continuous growth of the economy and the development of financial technology, the market scale of personal credit, as a financial tool to satisfy consumer demand, has naturally expanded. Influenced by the economic downward pressure, the increase of non-performing loan behaviors and various unexpected changes, the default rate of personal credit is gradually rising, and the importance of a perfect and efficient personal credit assessment model is self-evident. In the process of credit assessment, a series of specific indicators and factors are used to judge the credit risk of an individual, which requires a huge amount of resources under a huge market scale. In this paper, a logistic regression model based on sparse optimization is proposed, which can quickly produce individual risk assessment results while maintaining a certain degree of accuracy. Finally, the effectiveness of the proposed sparse logistic regression model is verified by real data.