计算机应用 ›› 2013, Vol. 33 ›› Issue (11): 3094-3096.

• 人工智能 • 上一篇    下一篇

基于稀疏贝叶斯学习的个人信用评估

李太勇1,2,王会军3,吴江2,3,张智林4,唐常杰5   

  1. 1. 西南财经大学
    2. 西南财经大学 中国支付体系研究中心,成都 610074;
    3. 西南财经大学 经济信息工程学院,成都 610074
    4. Emerging Technology Lab, Samsung Research and Development Institute America-Dallas, Richardson TX 75082, USA;
    5. 四川大学 计算机学院,成都 610064
  • 收稿日期:2013-05-17 修回日期:2013-07-16 出版日期:2013-11-01 发布日期:2013-12-04
  • 通讯作者: 李太勇
  • 作者简介:李太勇(1979-),男,四川安岳人,副教授,博士,CCF高级会员,主要研究方向:数据挖掘;王会军(1988-),男,山东潍坊人,硕士研究生,主要研究方向:金融风险;吴江(1980-),男,浙江衢州人,副教授,博士,主要研究方向:数据库与知识工程;张智林(1980-),男,湖南武陵人,博士,主要研究方向:稀疏贝叶斯学习;唐常杰(1946-),男,重庆人,教授,博士生导师,主要研究方向:数据库与知识工程。
  • 基金资助:
    教育部人文社会科学研究青年基金资助项目;中央高校基本科研业务专项资金资助项目;西南财经大学科研基金资助项目

Sparse Bayesian learning for credit risk evaluation

LI Taiyong1,2,WANG Huijun3,WU Jiang1,3,ZHANG Zhilin4,TANG Changjie5   

  1. 1. Institute of Chinese Payment System, Southwestern University of Finance and Economics, Chengdu Sichuan 610074, China;
    2. undefined
    3. School of Economic Information Engineering, Southwestern University of Finance and Economics, Chengdu Sichuan 610074, China;
    4. Emerging Technology Lab, Samsung Research and Development Institute America-Dallas, Richardson TX 75082, USA;
    5. School of Computer Science, Sichuan University, Chengdu Sichuan 610064, China
  • Received:2013-05-17 Revised:2013-07-16 Online:2013-12-04 Published:2013-11-01
  • Contact: LI Taiyong

摘要: 针对传统信用评估方法分类精度低、特征可解释性差等问题,提出了一种使用稀疏贝叶斯学习方法来进行个人信用评估的模型(SBLCredit)。SBLCredit充分利用稀疏贝叶斯学习的优势,在添加的特征权重的先验知识的情况下进行求解,使得特征权重尽量稀疏,以此实现个人信用评估和特征选择。在德国和澳大利亚真实信用数据集上,SBLCredit方法的分类精度比传统的K近邻、朴素贝叶斯、决策树和支持向量机平均提高了4.52%,6.40%,6.26%和2.27%。实验结果表明,SBLCredit分类精度高,选择的特征少,是一种有效的个人信用评估方法。

关键词: 稀疏贝叶斯学习, 分类, 信用评估, 金融风险, 特征选择

Abstract: To solve the low classification accuracy and poor interpretability of selected features in traditional credit risk evaluation, a new model using Sparse Bayesian Learning (SBL) to evaluate personal credit risk (SBLCredit) was proposed in this paper. The SBLCredit utilized the advantages of SBL to get as sparse as possible solutions under the priori knowledge on the weight of features, which led to both good classification performance and effective feature selection. SBLCredit improved the classification accuracy of 4.52%, 6.40%, 6.26% and 2.27% averagely when compared with the state-of-the-art K-Nearest Neighbour (KNN), Nave Bayes, decision tree and support vector machine respectively on real-world German and Australian credit datasets. The experimental results demonstrate that the proposed SBLCredit is a promising method for credit risk evaluation with higher accuracy and fewer features.

Key words: Sparse Bayesian Learning (SBL), classification, credit risk evaluation, financial risk, feature selection

中图分类号: