Implementation of the K-Means, K-Medoids, and Mini Batch K-Means Algorithms on the Classification of Natural Disaster Prone Areas in Indonesia

In recent years, Indonesia has experienced numerous natural disasters, as evidenced by data from the Indonesian Central Statistics Agency. Natural disasters are events or sequences of events that can pose significant threats to human life, disrupt daily activities, and lead to loss of life, property, and financial losses. This study focuses on identifying disaster-prone areas in Indonesia. The data, sourced from the Central Statistics Agency, covers natural disasters from 2018 to 2023 across 34 provinces. It includes information on earthquakes, tsunamis, volcano eruptions, landslides, floods, droughts, extreme weather, forest fires, and abrasions. The research employs unsupervised learning methods such as KMeans, K-Medoids, and Mini Batch K-Means for classification. Among these, the K-Means algorithm demonstrates the best performance, achieving a Davies-Bouldin Index (DBI) of 0.6303.