Integrating Multiple Data Sources and Learning Models to Predict Infectious Diseases in China

Area of Science:

• Epidemiology
• Public Health
• Data Science

Background:

• Infectious disease outbreaks pose significant risks to public health, societal well-being, and economic stability.
• Early warning systems are vital for mitigating the impact of disease outbreaks.
• Accurate prediction of infectious disease incidence is essential for effective public health interventions.

Purpose of the Study:

• To develop and evaluate predictive models for infectious disease morbidity in China.
• To compare the performance of various modeling techniques, including deep learning, for disease forecasting.
• To leverage historical incidence data, search engine queries, and seasonal information for enhanced prediction.

Main Methods:

• Collected historical morbidity and mortality data for 10 infectious diseases in China (2012-2016).
• Integrated search engine query data and seasonal information into prediction models.
• Constructed and compared linear, time series, boosting tree, and recurrent neural network (RNN) models.

Main Results:

• The recurrent neural network (RNN) model demonstrated superior predictive performance compared to other models.
• The integrated approach, combining historical data with search trends, improved forecasting accuracy.
• RNN models effectively predicted the incidence of various infectious diseases, including typhoid fever and tuberculosis.

Conclusions:

• Advanced prediction techniques, particularly RNNs, significantly enhance infectious disease forecasting capabilities.
• Integrating diverse data sources, such as search engine queries, improves the accuracy of early warning systems.
• Improved infectious disease prediction technologies can lead to more effective disease prevention and control strategies.