A data analytics approach for COVID-19 spread and end prediction (with a case study in Iran)

Area of Science:

• Epidemiology
• Data Science
• Public Health

Background:

• The COVID-19 pandemic significantly impacted Iran, with over 90,000 cases by April 30, 2020.
• Socio-economic challenges necessitate accurate outbreak trend prediction for effective management.

Purpose of the Study:

• To develop a beneficial method for predicting the COVID-19 outbreak trend in Iran.
• To enhance governmental awareness of key factors influencing the virus's future trajectory.

Main Methods:

• Compiled a dataset of confirmed cases, daily deaths, and recoveries.
• Integrated time-varying variables (behavior, policies) with machine learning algorithms (logistic function, Gaussian functions).
• Developed novel metrics: weekly death rate, life rate, mortality rate, and recovery rate.

Main Results:

• Gaussian functions demonstrated superior performance in predicting COVID-19 trends.
• The developed methods provide insights into factors impacting future outbreak patterns.
• Identified key variables influencing mortality and recovery rates.

Conclusions:

• Machine learning, particularly Gaussian functions, offers a robust approach for COVID-19 trend prediction.
• Accurate trend forecasting is crucial for informed public health policy and resource allocation.
• The study provides valuable insights for governmental awareness and strategic planning during the pandemic.