Comparison of K-Means and Two-Step Cluster Analysis Methods for Clustering COVID-19 Data

Abstract

This study compares the K-Means and two-step cluster analysis methods for clustering COVID-19 data. The dataset had 1,893,941 cumulative cases from January 2020 to October 2021. K-means clustering resulted in eight clusters, while two-step cluster analysis clustering resulted in three grouped cases by nationality, occupation, patient type, and risk group. These clusters were categorized based on age, gender, nationality, occupation, and region of infection. Group 1 had 5,883 workers infected in community settings, Group 2 had 7,420 foreign migrant workers infected in industrial settings or through direct contact with patients, and Group 3 had 6,870 cases of indirect transmission. The study recommends targeted interventions and continued monitoring and evaluation based on the clusters. The findings can help improve government policies, medical facilities, and treatment