A framework using large time series model for early warning of infectious diseases

Area of Science:

• Epidemiology
• Data Science
• Public Health

Background:

• Effective infectious disease control systems require early warning mechanisms to detect outbreaks.
• Existing anomaly detection methods struggle with data limitations in infectious disease surveillance.

Purpose of the Study:

• To develop an effective early warning framework for infectious diseases using generative pre-trained large time series models.
• To address the data quality and quantity constraints of current anomaly detection methods.

Main Methods:

• Proposed an early warning framework based on generative pre-trained large time series models.
• Evaluated the framework's performance against statistical and deep learning methods.

Main Results:

• The proposed framework demonstrated superior performance compared to traditional statistical and deep learning approaches.
• The framework requires less data while maintaining high accuracy in detecting infectious disease trends.

Conclusions:

• The developed framework offers a readily deployable solution for infectious disease early warning.
• This approach exhibits strong generalization capabilities and exceptional performance, benefiting epidemic modeling research.