Finding the Next Interesting Loan for Investors on a Peer-to-Peer Lending Platform

• Main Authors: Chichang Liu, Yidong Ming, Yingyuan Xiao, Wenguang Zheng, Ching-Hsien Hsu
• Format: Article
• Language: English
• Published: IEEE 2021-01-01
• Series: IEEE Access
• Subjects: P2P lending; recommender system; matrix factorization; deep learning
• Online Access: https://ieeexplore.ieee.org/document/9509424/

With the development of the mobile Internet, a peer-to-peer(P2P) online lending platform has become increasingly popular in the financial market, and it attracts a massive number of users. The task that helps investors find potential loans for improving the funding success rate has become a major challenge for lending platforms. However, the traditional recommendation schemes rarely take into account the challenges, such as the timeliness of loans (i.e., when a loan funding is completed or expired, it will no longer recruit investment), the common cold start problem (continuously releasing new loans is a common phenomenon), and the loans’ potential default risk. Considering the above characteristics, we propose a deep learning model based on a sequence of the incremental matrix factorization technology (DeepSeIMF). First, the cold start problem of loans can be effectively solved by designing an incremental matrix factorization model based on the time series. Then, a neural network is used to provide investors with personalized investment recommendation services based on risk assessment. Finally, the model performance is systematically evaluated based on a large-scale real-world dataset. The experimental results demonstrate the effectiveness of our solution.