Title: Peer-to-peer lending default prediction model: a credit scoring application with social media data

Authors: Taufik Faturohman; Sudarso Kaderi Wiryono; Hasna Laila Nabila Khilfah; Allesandra Andri; Muhammad Abdullah Hamzah; Okta Saputra; Gun Gun Indrayana

Addresses: School of Business and Management, Institut Teknologi Bandung, Bandung, West Java, 40132, Indonesia ' School of Business and Management, Institut Teknologi Bandung, Bandung, West Java, 40132, Indonesia ' School of Business and Management, Institut Teknologi Bandung, Bandung, West Java, 40132, Indonesia ' School of Business and Management, Institut Teknologi Bandung, Bandung, West Java, 40132, Indonesia ' School of Business and Management, Institut Teknologi Bandung, Bandung, West Java, 40132, Indonesia ' School of Business and Management, Institut Teknologi Bandung, Bandung, West Java, 40132, Indonesia ' School of Business and Management, Institut Teknologi Bandung, Bandung, West Java, 40132, Indonesia

Abstract: This study applied social media data to build credit scoring models in reducing non-performing loans in a peer-to-peer (P2P) lending portfolio. P2P lending can increase financial inclusion that has been known as one of the important factors in reducing poverty and enhancing prosperity. Discriminant analysis (DA), logistic regression (LR), neural network (NN), and support vector machine (SVM) were employed to compare effectiveness of the models. Results show that NN and SVM provide better outputs in predicting the default prediction model. Furthermore, social media data comprehensively increase the accuracy of lending default predictions.

Keywords: peer-to-peer lending; credit scoring; social media; neural network; SVM; support vector machine; financial technology; non-performing loan.

DOI: 10.1504/IJMEF.2024.139027

International Journal of Monetary Economics and Finance, 2024 Vol.17 No.2/3, pp.189 - 200

Received: 05 Aug 2022
Accepted: 12 Jan 2023
Published online: 10 Jun 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article