作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2020, Vol. 46 ›› Issue (10): 308-314. doi: 10.19678/j.issn.1000-3428.0056119

• 开发研究与工程应用 • 上一篇    下一篇

融合深度神经网络的个人信用评估方法

王重仁1, 王雯2,3, 佘杰4, 凌晨4,5   

  1. 1. 山东财经大学 管理科学与工程学院, 济南 250014;
    2. 济南大学 金融研究院, 济南 250001;
    3. 中泰证券股份有限公司风险管理部, 济南 250001;
    4. 上海财经大学 信息管理与工程学院, 上海 200433;
    5. 上海健康医学院 医疗器械学院, 上海 201318
  • 收稿日期:2019-09-25 修回日期:2019-11-12 发布日期:2019-11-19
  • 作者简介:王重仁(1984-),男,讲师、博士,主研方向为金融大数据分析;王雯,博士;佘杰,博士研究生;凌晨,讲师、博士研究生。
  • 基金资助:
    国家社会科学基金青年项目(19CJL041)。

Personal Credit Assessment Method Fused with Depth Neural Network

WANG Chongren1, WANG Wen2,3, SHE Jie4, LING Chen4,5   

  1. 1. School of Management Science and Engineering, Shandong University of Finance and Economics, Jinan 250014, China;
    2. Institute of Finance, Jinan University, Jinan 250001, China;
    3. Department of Risk Management, Zhongtai Securities Co., Ltd., Jinan 250001, China;
    4. School of Information Management and Engineering, Shanghai University of Finance and Economics, Shanghai 200433, China;
    5. School of Medical Devices, Shanghai University of Medicine & Health Sciences, Shanghai 201318, China
  • Received:2019-09-25 Revised:2019-11-12 Published:2019-11-19

摘要: 为提高信用风险评估的准确性,基于互联网行业的用户行为数据,提出一种基于长短期记忆(LSTM)神经网络和卷积神经网络(CNN)融合的深度神经网络个人信用评分方法。对每个用户的行为数据进行编码,形成一个包括时间维度和行为维度的矩阵,通过融合基于注意力机制的LSTM模型和CNN模型2个子模型,从用户原始行为数据中提取序列特征和局部特征。在真实数据集上的实验结果表明,该方法的KS指标和AUC指标均优于传统的机器学习方法和单一的LSTM卷积神经网络方法,证明了该方法在个人信用评分领域的有效性和可行性。

关键词: 大数据, 个人信用评分, 机器学习, 深度神经网络, 卷积神经网络, 长短期记忆神经网络

Abstract: To improve the accuracy of credit risk assessment,based on the user behavior data of the Internet industry,this paper proposes a personal credit scoring method based on fused deep neural network combining Long Short-Term Memory(LSTM) neural network and Convolutional Neural Network(CNN).The behavior data of each user is encoded to form a matrix that includes the time dimension and the behavior dimension.By fusing the two sub-models,LSTM model and CNN model based on the attention mechanism,the sequence features and local features are extracted from the original user behavior data.Experimental results on real datasets show that the proposed method outperforms the traditional machine learning methods and the single LSTM convolutional neural network method in terms of KS index and AUC index,demonstrating the effectiveness and feasibility of this method in the field of personal credit scoring.

Key words: big data, personal credit scoring, machine learning, deep neural network, Convolutional Neural Network(CNN), Long and Short-Term Memory (LSTM) neural network

中图分类号: