Community detection using unsupervised machine learning techniques on COVID-19 dataset.

Area of Science:

• Epidemiology
• Data Science
• Public Health

Background:

• The COVID-19 pandemic has had unprecedented global impacts across economic, medical, and sociological domains.
• Extensive research efforts are underway to understand and mitigate the effects of this global health crisis.
• Analyzing epidemiological data is crucial for developing effective response strategies.

Purpose of the Study:

• To analyze COVID-19 trends, regional impact, and country-level case variations.
• To identify significant variables influencing the pandemic using dimensionality reduction.
• To uncover community structures among countries for targeted interventions.

Main Methods:

• Principal Component Analysis (PCA) was employed for dimensionality reduction and identification of key variables.
• Unsupervised K-means clustering was applied to group countries based on pandemic data.
• Comparative analysis of clustering results with and without PCA pre-processing.

Main Results:

• PCA effectively reduced data dimensionality and highlighted significant COVID-19 related variables.
• K-means clustering, following PCA, identified more precise community structures among affected countries.
• The identified country communities offer insights into regional pandemic dynamics.

Conclusions:

• PCA-driven K-means clustering provides a more accurate method for community detection in epidemiological datasets.
• The identified country groupings can inform tailored policy-making and health sector management strategies.
• This approach aids researchers and policymakers in understanding and addressing the multifaceted challenges of the COVID-19 pandemic.