A Novel behavioral scoring model for estimating probability of default over time in peer-to-peer lending
作者：
Highlights：
• We propose a novel behavioral scoring model for online peer-to-peer lending.
• The proposed model predicts if and when borrowers will default.
• For incidence component, random forest is used to predict whether default.
• For latency component, random survival forest is used to predict when to default.
• The empirical study is implemented on a real-world dataset form P2P platform.
摘要
•We propose a novel behavioral scoring model for online peer-to-peer lending.•The proposed model predicts if and when borrowers will default.•For incidence component, random forest is used to predict whether default.•For latency component, random survival forest is used to predict when to default.•The empirical study is implemented on a real-world dataset form P2P platform.
论文关键词：Behavioral scoring,Dynamic probability,P2P lending,Random forest,Random survival forest,Risk management,Survival analysis
论文评审过程：Received 12 September 2017, Revised 18 December 2017, Accepted 18 December 2017, Available online 21 December 2017, Version of Record 4 January 2018.
论文官网地址：https://doi.org/10.1016/j.elerap.2017.12.006
