Kalman filter based short term prediction model for COVID-19 spread.

Area of Science:

• Epidemiology
• Data Science
• Computational Biology

Background:

• The Corona Virus Disease 2019 (COVID-19) pandemic presents a significant global health challenge.
• Understanding the dynamics of SARS-CoV-2 spread, including influencing factors, is crucial for effective public health interventions.

Purpose of the Study:

• To analyze the spread of COVID-19 using machine learning techniques.
• To identify key demographic and environmental factors contributing to SARS-CoV-2 transmission.
• To forecast short-term and long-term spread patterns of the virus.

Main Methods:

• Data integration from various sources on COVID-19 spread.
• Application of Machine Learning models: Random Forest for factor identification and feature importance analysis.
• Utilizing Pearson Correlation matrix for visualizing linear relationships between features.
• Employing Kalman Filter for short-term and long-term spread forecasting.

Main Results:

• Random Forest model demonstrated strong performance in evaluating COVID-19 spread data.
• Key demographic and environmental factors influencing virus transmission were identified and their contributions analyzed.
• Pearson Correlation heatmap provided insights into feature relationships.
• Kalman Filter showed satisfactory short-term forecasting accuracy but limited long-term predictive power.

Conclusions:

• Machine learning techniques, particularly Random Forest and Kalman Filter, are effective tools for analyzing and understanding COVID-19 spread dynamics.
• The study successfully identified significant contributing factors to the pandemic's progression.
• While Kalman Filter is useful for short-term forecasting, further research is needed for robust long-term predictions.