Time series predicting of COVID-19 based on deep learning

Area of Science:

• Epidemiology
• Computational Biology
• Data Science

Background:

• The COVID-19 pandemic, declared by WHO in March 2020, prompted extensive research into outbreak prediction.
• Time-series variables are recognized as key factors influencing the spread of infectious diseases like influenza and SARS.
• Accurate forecasting models are crucial for public health response and resource allocation during pandemics.

Purpose of the Study:

• To develop and assess deep learning models for predicting COVID-19 outbreak trajectories in Malaysia, Morocco, and Saudi Arabia.
• To compare the predictive performance of Recurrent Neural Networks (RNNs) and Long Short-Term Memory (LSTM) networks for COVID-19 case numbers.
• To forecast confirmed COVID-19 cases and deaths for a seven-day period based on available data.

Main Methods:

• Utilized public datasets from the European Centre for Disease Prevention and Control.
• Applied deep learning models, specifically RNN and LSTM networks, for outbreak prediction.
• Assessed key features influencing COVID-19 trends and compared case and death numbers across the three countries.

Main Results:

• LSTM models demonstrated a high precision accuracy of 98.58% in predicting COVID-19 cases.
• RNN models achieved a precision accuracy of 93.45%.
• The study successfully predicted confirmed COVID-19 cases and fatalities for a subsequent seven-day period.

Conclusions:

• Deep learning, particularly LSTM networks, offers a highly accurate approach for predicting COVID-19 spread and impact.
• The developed models provide valuable insights for public health planning and intervention strategies in affected regions.
• Accurate short-term forecasting aids in managing resources and mitigating the effects of the pandemic.