Forecasting peer-to-peer platform default rate with LSTM neural network
作者：
Highlights：
• Monthly default rate is essential for revealing the health of a P2P lending platform.
• LSTM models are proposed for P2P lending platform monthly default rate forecast.
• LSTM models can extract and memorize the time series information of default rate.
• Models’ comparation is under two cross-validation methods and three time-dimensions.
• LSTM outperforms several benchmark methods in both prediction and trend accuracy.
摘要
•Monthly default rate is essential for revealing the health of a P2P lending platform.•LSTM models are proposed for P2P lending platform monthly default rate forecast.•LSTM models can extract and memorize the time series information of default rate.•Models’ comparation is under two cross-validation methods and three time-dimensions.•LSTM outperforms several benchmark methods in both prediction and trend accuracy.
论文关键词：Peer-to-peer lending,Credit default,LSTM neural network,Time-series prediction
论文评审过程：Received 20 November 2019, Revised 5 May 2020, Accepted 11 May 2020, Available online 17 July 2020, Version of Record 6 August 2020.
论文官网地址：https://doi.org/10.1016/j.elerap.2020.100997
