计算机科学 ›› 2021, Vol. 48 ›› Issue (11A): 429-434.doi: 10.11896/jsjkx.201000013
王萧萧1, 王亭雯1, 马玉玲2, 范佳奕3, 崔超然1
WANG Xiao-xiao1, WANG Ting-wen1, MA Yu-ling2, FAN Jia-yi3, CUI Chao-ran1
摘要： P2P网络借贷是近年来新兴的一种金融业务模式,具有投资门槛低、交易方便快捷、融资成本低等优点。但在快速成长的同时,借贷过程中的信用风险问题也日益凸显,层出不穷的借款人跑路乃至诈骗事件给行业留下重大阴影。针对该问题,提出一种基于深度森林的网贷借款人信用风险评估方法。首先从借款人的基本信息和历史借款信息两类数据中提取特征;然后通过多粒度扫描和级联森林模块构建深度森林模型,对借款人进行违约预测,同时使用基尼指数计算随机森林的特征重要性评分,并使用波达计数法进行排序融合,从而对模型的预测结果给出一定的解释。在LendingClub和拍拍贷两个公开数据集上,将所提出的方法与支持向量机、随机森林和广而深的网络等方法进行了对比,实验表明该方法具有更好的性能,并且特征重要性评分符合人们的直观理解和客观认识。
中图分类号:
| [1]OHLSON J A.Financial Ratios and the Probabilistic Prediction of Bankruptcy[J].Journal of Accounting Research,1980,18(1):109-131. [2]XIAO H M,HOU Y,CUI C N.Evaluation of P2P Lending Borrower's Credit on BP Artificial Neural Network [J].Operations Research and Management,2018,27 (9):112-118. [3]ZHOU Z H,FENG J.Deep Forest:Towards An Alternative to Deep Neural Networks[C]//Proceedings of the 26th International Joint Conference on Artificial Intelligence.2017:3553-3559. [4]BREIMAN L,FRIEDMAN J,OLSHEN R,et al.Classificationand Regression Trees[M].New York:Chapman & Hall,1984. [5]LU H Y.Construction of risk evaluation system of P2P online loan platform based on SVM [J].Science and Technology Economics Market,2018(2):70-74. [6]TAN Z M,XIE K,PENG Y P.Research on Credit Risk Evalua- tion of P2P Online Borrowers Based on Gradient Boosting Decision Tree Model [J].Soft Science,2018,32(12):136-140. [7]XU T T.Application of random forest in credit risk assessment of P2P online loan borrowing [D].Jinan:Shandong University,2017. [8]MA P J,WANG Y,YU L,et al.Risk assessment of P2P net-work lending based on cost-sensitive decision tree [J].Computer Integrated Manufacturing System,2018,243 (7):296-302. [9]ZHANG Y C,SONG X P,LUO Y.Research on Customer CreditEvaluation Based on Fuzzy Support Vector Machine [J].Statistics and Decision,2008(7):16-19. [10]WANG C R,HAN D M,LIU Q G,et al.A Deep Learning Approach for Credit Scoring of Peer-to-Peer Lending Using Attention Mechanism LSTM[J].IEEE Access,2018(7):2161-2168. [11]YANG Z,ZHANG Y S,GUO B H,et al.DeepCredit:Exploiting User Cickstream for Loan Risk Prediction in P2P Lending[C]//International AAAI Conference on Web and Social Media Twelfth International AAAI Conference on Web and Social Media.Palo Alto,California USA:AAAI,2018:444-443. [12]BASTANI K,ASGARI E,NAMAVARI H.Wide and deeplearning for peer-to-peer lending[J].Expert Systems With Applications,2019,134:209-224. [13]TONG T,LUO S L,PAN L M,Zhang Tiemei.Scale data mining method based on deep forest[J].Electronic Design Engineering,2020,28(13):88-91,96. [14]UTKIN L V,RYABININ M A.A Siamese Deep Forest[J].Journal of Knowledge-Based Systems[J].arXiv:1704.08715vl,2017:5-6. [15]GE S L,YE J,HE M X.Prediction model of user purchase behavior based on deep forest[J].Computer Science,2019,46(9):190-194. [16]LU X D,DUAN Z M,QIAN Y K,et al.A Malicious Code Classification Method Based on Deep Forest[J].Journal of Software,2020,31(5):1454-1464. |
| [1] | 蒋鹏飞, 魏松杰. 基于深度森林与CWGAN-GP的移动应用网络行为分类与评估 Classification and Evaluation of Mobile Application Network Behavior Based on Deep Forest and CWGAN-GP 计算机科学, 2020, 47(1): 287-292. https://doi.org/10.11896/jsjkx.181102118 |
| [2] | 葛绍林, 叶剑, 何明祥. 基于深度森林的用户购买行为预测模型 Prediction Model of User Purchase Behavior Based on Deep Forest 计算机科学, 2019, 46(9): 190-194. https://doi.org/10.11896/j.issn.1002-137X.2019.09.027 |
| [3] | 韩慧,王黎明,柴玉梅,刘箴. 基于强化表征学习深度森林的文本情感分类 Text Sentiment Classification Based on Deep Forests with Enhanced Features 计算机科学, 2019, 46(7): 172-179. https://doi.org/10.11896/j.issn.1002-137X.2019.07.027 |
| [4] | 杨德杰, 章宁, 袁戟, 白璐. 基于堆栈降噪自编码网络的个人信用风险评估方法 Individual Credit Risk Assessment Based on Stacked Denoising Autoencoder Networks 计算机科学, 2019, 46(10): 7-13. https://doi.org/10.11896/jsjkx.181102216 |
| [5] | 薛参观, 燕雪峰. 基于改进深度森林算法的软件缺陷预测 Software Defect Prediction Based on Improved Deep Forest Algorithm 计算机科学, 2018, 45(8): 160-165. https://doi.org/10.11896/j.issn.1002-137X.2018.08.029 |
| [6] | 刘盼,李华康,孙国梓. 基于短时多源回归算法的P2P平台风险观测方法 Risk Observing Method Based on Short-time Multi-source Regression Algorithm on P2P Platform 计算机科学, 2018, 45(5): 97-101. https://doi.org/10.11896/j.issn.1002-137X.2018.05.017 |
| [7] | 任永功,杨雪,杨荣杰,胡志冬. 基于信息增益特征关联树的文本特征选择算法 Text Feature Selection Methods Based on Information Gain and Feature Relation Tree 计算机科学, 2013, 40(10): 252-256. |
| [8] | 任永功 杨荣杰 尹明飞 马名威. 基于信息增益的文本特征选择方法 Information-gain-based Text Feature Selection Method 计算机科学, 2012, 39(11): 127-130. |
|
||
首页