The Application of Machine Learning Algorithms in Predicting the Borrower’s Default Risk in Online Peer-to-Peer Lending
Download as PDF
DOI: 10.25236/icemit.2019.016
Author(s)
Huixi He, Kazumitsu Nawata
Corresponding Author
Huixi He
Abstract
The machine learning algorithms have excellent performance in classification, regression, clustering, and visualization, which should also be applicated in finance studies such as bank failure prediction, risk management, and so on. This paper focuses on the application of machine learning techniques in predicting the default risk of an individual borrower in a P2P lending platform based on data from one of the most famous platforms. We use four widely accepted classification algorithms to evaluate the loan risk of a certain borrower by predicting whether this borrower will overdue. In the data processing step, we highlight our work in applying LDA in text analytics. We found that logistic regression, Random Forest, Neural Networks, and Naive Bayes all performed well with the precision above 86%, and Naive Bayes reaches 97%.
Keywords
Logistic regression, random forest, neural networks, naive bayes, P2P lending, default risk