Learning Short-Term and Long-Term Patterns of High-Order Dynamics in Real-World Networks

TL;DR

This paper introduces LINCOLN, a novel method for modeling high-order, dynamic relationships in real-world networks by capturing both short-term influences and long-term periodic patterns, leading to improved prediction accuracy.

Contribution

LINCOLN is the first approach to simultaneously model short-term and long-term high-order dynamics in real-world networks using hyperedge encoding and periodic pattern integration.

Findings

LINCOLN outperforms nine state-of-the-art methods in dynamic hyperedge prediction.

The method effectively captures both immediate and periodic long-term network patterns.

Experimental results demonstrate significant improvements in prediction accuracy.

Abstract

Real-world networks have high-order relationships among objects and they evolve over time. To capture such dynamics, many works have been studied in a range of fields. Via an in-depth preliminary analysis, we observe two important characteristics of high-order dynamics in real-world networks: high-order relations tend to (O1) have a structural and temporal influence on other relations in a short term and (O2) periodically re-appear in a long term. In this paper, we propose LINCOLN, a method for Learning hIgh-order dyNamiCs Of reaL-world Networks, that employs (1) bi-interactional hyperedge encoding for short-term patterns, (2) periodic time injection and (3) intermediate node representation for long-term patterns. Via extensive experiments, we show that LINCOLN outperforms nine state-of-the-art methods in the dynamic hyperedge prediction task.