Attention-based Dynamic Multilayer Graph Neural Networks for Loan Default Prediction

TL;DR

This paper introduces a dynamic multilayer graph neural network model with attention mechanisms for improved loan default prediction, capturing evolving borrower connections over time.

Contribution

It presents a novel combination of GNN, RNN, and attention mechanisms to model dynamic, multilayer borrower networks for credit risk assessment.

Findings

Enhanced default prediction accuracy over traditional methods

Attention mechanism identifies key time periods and connections

Model provides insights into the influence of borrower networks

Abstract

Whereas traditional credit scoring tends to employ only individual borrower- or loan-level predictors, it has been acknowledged for some time that connections between borrowers may result in default risk propagating over a network. In this paper, we present a model for credit risk assessment leveraging a dynamic multilayer network built from a Graph Neural Network and a Recurrent Neural Network, each layer reflecting a different source of network connection. We test our methodology in a behavioural credit scoring context using a dataset provided by U.S. mortgage financier Freddie Mac, in which different types of connections arise from the geographical location of the borrower and their choice of mortgage provider. The proposed model considers both types of connections and the evolution of these connections over time. We enhance the model by using a custom attention mechanism that weights the different time snapshots according to their importance. After testing multiple configurations, a model with GAT, LSTM, and the attention mechanism provides the best results. Empirical results demonstrate that, when it comes to predicting probability of default for the borrowers, our proposed model brings both better results and novel insights for the analysis of the importance of connections and timestamps, compared to traditional methods.