A decision support model for investment on P2P lending platform

Area of Science:

• * Financial Technology (FinTech)
• * Computational Economics
• * Data Science

Background:

• * Peer-to-peer (P2P) lending presents challenges in making effective investment decisions due to complex interconnections.
• * Traditional P2P platforms require robust models to evaluate loan risks and investment opportunities.
• * The bipartite graph structure inherent in P2P lending platforms offers a unique modeling approach.

Purpose of the Study:

• * To develop and validate an iterative computation model for evaluating unknown loans in P2P lending.
• * To assess the efficacy of the proposed model in aiding both borrowers and lenders in P2P platforms.
• * To explore the synergistic benefits of integrating the iterative model with Logistic Regression for improved performance.

Main Methods:

• * Development of an iterative computation model based on the bipartite graph structure of P2P lending.
• * Extensive experimentation using real-world data from the Prosper P2P lending marketplace.
• * Comparative analysis against Bayes and Logistic Regression models, followed by the creation and testing of a hybrid model.

Main Results:

• * The proposed iterative computation model effectively aids borrowers in selecting suitable loans and lenders in making sound investment decisions.
• * Logistic Regression demonstrated a complementary role to the iterative computation model.
• * The hybrid model, integrating iterative computation and Logistic Regression, exhibited superior efficiency and stability compared to individual models.

Conclusions:

• * The iterative computation model provides a valuable tool for enhancing decision-making in P2P lending.
• * A hybrid approach combining iterative computation with Logistic Regression offers significant advantages for P2P lending platforms.
• * The developed hybrid model is more efficient and stable than standalone methods for P2P investment analysis.