MGLEP: Multimodal Graph Learning for Modeling Emerging Pandemics with Big Data.

Area of Science:

• Epidemiology
• Data Science
• Public Health

Background:

• Accurate pandemic forecasting is vital for public health.
• Traditional methods often miss early indicators by relying solely on epidemiological data.

Purpose of the Study:

• To propose a novel framework, MGLEP, for enhanced pandemic forecasting and analysis.
• To integrate multi-modal data, including social media, with temporal graph neural networks.

Main Methods:

• Developed the MGLEP framework integrating temporal graph neural networks and multi-modal data.
• Utilized pre-trained language models for social media content analysis.
• Discovered user interaction graph structures to identify pandemic patterns.

Main Results:

• MGLEP demonstrated superior performance in pandemic forecasting and analysis compared to baseline methods.
• The framework showed effectiveness across diverse pandemic scenarios and prediction timelines.
• Achieved comprehensive pandemic landscape understanding with reduced time lag and cost.

Conclusions:

• The fusion of temporal graph learning and multi-modal data offers a powerful approach to pandemic surveillance.
• MGLEP provides richer, more timely indicators for effective public health decision-making.
• This integrated framework enhances the ability to predict and manage emerging infectious disease threats.