Exploring the Discretization of Peak Ground Acceleration Variables: a Comparative Analysis of K-means and K-medoids

The susceptibility of Indonesia to seismic activity as a result of intricate interactions between tectonic plates highlights the imperative for a robust policy aimed at mitigating the impact of such events. Although accurately predicting the precise timing of an earthquake remains a formidable task, evaluating the possibility for resultant damage can significantly enhance efforts in disaster preparedness. The primary objective of this study is to investigate the process of discretizing the variable known as peak ground acceleration (PGA). Discretization refers to the process of transforming a continuous variable into a finite set of discrete values. In the present context, the K-means and K-medoids clustering algorithms were modified and assessed in terms of their effectiveness in discretizing PGA variables. The utilized assessment metric is the average silhouette width, which serves as a quantifiable indicator of the quality of clusters. Based on a thorough analysis and comparison of outcomes, it has been ascertained that the K-means algorithm has enhanced discretization capabilities, leading to the formation of distinct clusters characterized by increased cohesiveness and separation. The aforementioned findings provide a contribution to the progress of earthquake damage assessment approaches and offer vital insights for disaster management and response planning in regions prone to earthquakes. The correlation between the discretization capabilities of the K-means algorithm and its contribution to damage assessment, particularly considering that PGA is a critical factor in determining damage levels. PGA indicates the strength and duration of the shock depending on the size and location of the source. Information from the PGAcan be useful for minimizing losses arising from earthquake phenomena in the area. Therefore, studying the magnitude of PGA, one of which is maximum ground acceleration, is important in evaluating the impact it has on the condition of an area and can be a reference for the government and society in the process of building earthquake-resistant structural and infrastructure facilities.