Data Analytics for Predicting COVID-19 Cases in Top Affected Countries: Observations and Recommendations.

Area of Science:

• Epidemiology
• Data Science
• Artificial Intelligence

Background:

• The COVID-19 pandemic overwhelmed healthcare systems globally, necessitating accurate case number predictions.
• Existing prediction models often neglect external factors influencing virus spread, leading to inaccuracies.
• Effective forecasting is vital for governments to implement timely public health interventions.

Purpose of the Study:

• To develop an improved COVID-19 case prediction model by integrating historical data with external influencing factors.
• To enhance the accuracy of forecasting the number of COVID-19 cases for better resource allocation and policy-making.

Main Methods:

• Development of a nonlinear autoregressive exogenous input (NARX) neural network-based algorithm.
• Utilizing data analytics to incorporate both time-series case data and external variables.
• Experimental validation using data from highly affected countries across different continents.

Main Results:

• The developed NARX model demonstrated superior accuracy in predicting COVID-19 cases compared to existing methods.
• The algorithm successfully integrated historical trends and external factors for more reliable forecasts.
• Future predictions for August-September 2020 indicated the model's utility for proactive planning.

Conclusions:

• The proposed data analytics approach, particularly the NARX model, offers a significant improvement in COVID-19 case prediction accuracy.
• Accurate forecasting enables governments and the public to prepare and implement measures for resuming normal activities.
• This method provides a valuable tool for managing the ongoing pandemic and future public health crises.