A new integrated similarity measure for enhancing instance-based credit assessment in P2P lending

作者：

Highlights：

• A new integrated loan similarity measure combines lender and borrower information.

• A novel and general machine learning framework to integrate multiple similarities.

• Identifying conflicts between different similarity measures by non-order-preserving.

• The integrated measure effectively improves instance-based credit assessment method.

摘要

•A new integrated loan similarity measure combines lender and borrower information.•A novel and general machine learning framework to integrate multiple similarities.•Identifying conflicts between different similarity measures by non-order-preserving.•The integrated measure effectively improves instance-based credit assessment method.

论文关键词：P2P lending,Credit assessment,Instance-based learning,Loan similarity,Borrower and lender

论文评审过程：Received 29 June 2020, Revised 31 December 2020, Accepted 25 February 2021, Available online 4 March 2021, Version of Record 27 March 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.114798