Cost-sensitive multiple-instance learning method with dynamic transactional data for personal credit scoring

作者：

Highlights：

• A machine learning methodology is proposed for analyzing and scoring personal credit.

• Cost-sensitive multiple-instance learning is developed to extract credit features from dynamic transactional data.

• Transaction features are combined with personal and application features for feature engineering.

• Our research contributes to advance the computational method for credit scoring.

摘要

•A machine learning methodology is proposed for analyzing and scoring personal credit.•Cost-sensitive multiple-instance learning is developed to extract credit features from dynamic transactional data.•Transaction features are combined with personal and application features for feature engineering.•Our research contributes to advance the computational method for credit scoring.

论文关键词：Credit risk assessment,Dynamic transactional data,Cost-sensitive learning,Multiple-instance learning

论文评审过程：Received 29 September 2019, Revised 26 April 2020, Accepted 26 April 2020, Available online 5 May 2020, Version of Record 11 May 2020.

论文官网地址：https://doi.org/10.1016/j.eswa.2020.113489