ABSTRACT
The 2019 novel coronavirus disease (COVID-19) pandemic in Indonesia has caused issues in many sectors such as health, economy, and education. Several actions had been taken by the government to prevent and forestall the spread of the coronavirus infection. However, right now there are still many new cases emerging especially in cities with dense population. In the meantime, actions taken from the government are based on the classification of the severity of new cases;there are red zone, yellow zone and green zone. Therefore, mapping cities into zone is critical because it concerns the right decision to be implemented. This paper aimed to cluster the severity of each province in Indonesia based on the number of cases, recovered, and casualties using 3 clustering methods namely K-Means, K-Medoids, and Gaussian mixture model. The result shows that the most optimal clustering method is the Gaussian mixture model, while the least optimal method for clustering is the K-Means. Furthermore, it is also discovered that the cluster always changes overtime, and the cluster can shift depending on the corresponding parameter. © 2022 ICIC International.