A Comprehensive Investigation: Machine Learning for P2P Lending Prediction

Qichang Ma

2024

Abstract

Peer-to-Peer (P2P) lending, a new business format, has developed rapidly. However, there have also been many frauds, which have brought great challenges to investors and P2P lending platforms. This paper reviews the methods of P2P default prediction based on machine learning. Four common machine learning models are introduced: decision tree, support vector machine, deep neural networks and ensemble learning. For each model, the entire process of constructing the model in the literature is illustrated to describe how these models are applied in P2P default prediction. Three main challenges in using machine learning for P2P default prediction are proposed: because of the "black box" property in machine learning methods, P2P default prediction faces difficulties in interpretability; different models use data of different national sources, structures, features, etc., existing models cannot be directly applied to a specific P2P lending platform, so P2P default prediction faces difficulties in applicability; P2P default prediction involves a large amount of important user privacy and security data, so P2P default prediction also faces difficulties in privacy. Solutions to these difficulties are also proposed. This article provides a good review of using machine models for P2P default prediction, which can provide inspiration for managers of P2P platforms.

Download


Paper Citation


in Harvard Style

Ma Q. (2024). A Comprehensive Investigation: Machine Learning for P2P Lending Prediction. In Proceedings of the 1st International Conference on E-commerce and Artificial Intelligence - Volume 1: ECAI; ISBN 978-989-758-726-9, SciTePress, pages 66-70. DOI: 10.5220/0013206700004568


in Bibtex Style

@conference{ecai24,
author={Qichang Ma},
title={A Comprehensive Investigation: Machine Learning for P2P Lending Prediction},
booktitle={Proceedings of the 1st International Conference on E-commerce and Artificial Intelligence - Volume 1: ECAI},
year={2024},
pages={66-70},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013206700004568},
isbn={978-989-758-726-9},
}


in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on E-commerce and Artificial Intelligence - Volume 1: ECAI
TI - A Comprehensive Investigation: Machine Learning for P2P Lending Prediction
SN - 978-989-758-726-9
AU - Ma Q.
PY - 2024
SP - 66
EP - 70
DO - 10.5220/0013206700004568
PB - SciTePress