Deep learning in public health: Comparative predictive models for COVID-19 case forecasting.

Area of Science:

• Epidemiology
• Data Science
• Public Health

Background:

• The COVID-19 pandemic necessitated robust forecasting for public health policy in the UAE and Malaysia.
• Accurate prediction of infectious disease spread is critical for effective pandemic response.

Purpose of the Study:

• To compare the efficacy of various deep learning models for forecasting COVID-19 cases in the UAE and Malaysia.
• To identify the most suitable deep learning model architectures for these specific regions.

Main Methods:

• Evaluated Long Short-Term Memory (LSTM), Convolutional Neural Networks (CNN), and other deep learning models.
• Utilized confirmed case data, demographics, and socioeconomic factors, enhanced by Bayesian optimization.
• Employed predictive and retrospective analytical approaches for data interpretation.

Main Results:

• Deep learning algorithms demonstrated proficiency in forecasting COVID-19 cases, with varying effectiveness across models.
• Specific model architectures were identified as most suitable for the UAE and Malaysia's unique pandemic conditions.
• Bayesian optimization improved model performance in predicting case trajectories.

Conclusions:

• Deep learning models are valuable tools for precise and timely COVID-19 case forecasting.
• Findings provide crucial insights for developing targeted public health interventions in the UAE and Malaysia.
• The study highlights the utility of deep learning in processing complex health data for reliable projections.