The University of Southampton
University of Southampton Institutional Repository

How can lenders prosper? Comparing machine learning approaches to identify profitable peer-to-peer loan investments

How can lenders prosper? Comparing machine learning approaches to identify profitable peer-to-peer loan investments
How can lenders prosper? Comparing machine learning approaches to identify profitable peer-to-peer loan investments

Successful Peer-to-Peer (P2P) lending requires an evaluation of loan profitability from a large universe of loans. Predictions of loan profitability may be useful to rank potential investments. We investigate whether various types of prediction methods and the types of information contained in loan listing features matter for profitable investment. A range of methods and performance metrics are used to benchmark predictive performance, based on a large dataset of P2P loans issued on Lending Club. Robust linear mixed models are used to investigate performance differences between models, according to whether they assume linearity, whether they build ensembles, and which types of predictors they use. The main findings are that: linear methods perform surprisingly well on several (but not all) criteria; whether ensemble methods perform better than individual methods is measure dependent; the use of alternative text-based information does not improve profit scoring outcomes. We conclude that P2P lenders could potentially increase their investment returns by applying linear methods that directly predict the internal rate of return instead of other dependent variables such as loan default.

Credit scoring, Ensemble learning, Investment analysis, P2P Lending, Predictive modelling
0377-2217
1-12
Fitzpatrick, Trevor
b3d78774-8c4d-4f7d-875c-8483843da9ef
Mues, Christophe
07438e46-bad6-48ba-8f56-f945bc2ff934
Fitzpatrick, Trevor
b3d78774-8c4d-4f7d-875c-8483843da9ef
Mues, Christophe
07438e46-bad6-48ba-8f56-f945bc2ff934

Fitzpatrick, Trevor and Mues, Christophe (2021) How can lenders prosper? Comparing machine learning approaches to identify profitable peer-to-peer loan investments. European Journal of Operational Research, 294 (2), 1-12. (doi:10.1016/j.ejor.2021.01.047).

Record type: Article

Abstract

Successful Peer-to-Peer (P2P) lending requires an evaluation of loan profitability from a large universe of loans. Predictions of loan profitability may be useful to rank potential investments. We investigate whether various types of prediction methods and the types of information contained in loan listing features matter for profitable investment. A range of methods and performance metrics are used to benchmark predictive performance, based on a large dataset of P2P loans issued on Lending Club. Robust linear mixed models are used to investigate performance differences between models, according to whether they assume linearity, whether they build ensembles, and which types of predictors they use. The main findings are that: linear methods perform surprisingly well on several (but not all) criteria; whether ensemble methods perform better than individual methods is measure dependent; the use of alternative text-based information does not improve profit scoring outcomes. We conclude that P2P lenders could potentially increase their investment returns by applying linear methods that directly predict the internal rate of return instead of other dependent variables such as loan default.

Text
EJOR-D-19-00396_R3 - Accepted Manuscript
Restricted to Repository staff only until 10 February 2023.
Request a copy

More information

Accepted/In Press date: 30 January 2021
e-pub ahead of print date: 10 February 2021
Keywords: Credit scoring, Ensemble learning, Investment analysis, P2P Lending, Predictive modelling

Identifiers

Local EPrints ID: 447873
URI: http://eprints.soton.ac.uk/id/eprint/447873
ISSN: 0377-2217
PURE UUID: d1272d6c-7ddc-4a3b-93fc-17591e0084e8
ORCID for Christophe Mues: ORCID iD orcid.org/0000-0002-6289-5490

Catalogue record

Date deposited: 25 Mar 2021 18:18
Last modified: 28 Apr 2022 01:53

Export record

Altmetrics

Contributors

Author: Trevor Fitzpatrick
Author: Christophe Mues ORCID iD

Download statistics

Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.

View more statistics

Atom RSS 1.0 RSS 2.0

Contact ePrints Soton: eprints@soton.ac.uk

ePrints Soton supports OAI 2.0 with a base URL of http://eprints.soton.ac.uk/cgi/oai2

This repository has been built using EPrints software, developed at the University of Southampton, but available to everyone to use.

We use cookies to ensure that we give you the best experience on our website. If you continue without changing your settings, we will assume that you are happy to receive cookies on the University of Southampton website.

×