Dynamic Link and Flow Prediction in Bank Transfer Networks

TL;DR

This paper presents a novel dynamic prediction model for both link existence and weight in complex bank transfer networks, utilizing self-attention mechanisms and separate sub-tasks, validated on real-world datasets.

Contribution

Introduces a new model that predicts link presence and weight simultaneously using a two-step approach with self-attention, addressing scalability and complexity issues.

Findings

Effective prediction of link existence and weight in real-world networks

Outperforms traditional models in large-scale, sparse networks

Validated on cryptocurrency and bank transfer datasets

Abstract

The prediction of both the existence and weight of network links at future time points is essential as complex networks evolve over time. Traditional methods, such as vector autoregression and factor models, have been applied to small, dense networks, but become computationally impractical for large-scale, sparse, and complex networks. Some machine learning models address dynamic link prediction, but few address the simultaneous prediction of both link presence and weight. Therefore, we introduce a novel model that dynamically predicts link presence and weight by dividing the task into two sub-tasks: predicting remittance ratios and forecasting the total remittance volume. We use a self-attention mechanism that combines temporal-topological neighborhood features to predict remittance ratios and use a separate model to forecast the total remittance volume. We achieve the final prediction by multiplying the outputs of these models. We validated our approach using two real-world datasets: a cryptocurrency network and bank transfer network.