OPTIMIZATION MODEL IN CLUSTERING THE HAZARD ZONE AFTER AN EARTHQUAKE DISASTER

Abstract

There are a large number of approaches to clustering problems, including optimization-based methods involving mathematical programming models to develop efficient and meaningful clustering schemes. Clustering is one of the data labeling techniques. K-means clustering is a partition clustering algorithm that starts by selecting k representative points as the initial centroid. Each point is then assigned to the nearest centroid based on the selected specific proximity measure. This writing is focused on the grouping of post-earthquake hazard zones based on grouping with regard to certain characteristics which aim to describe the process of partitioning the N-dimensional population into K-sets based on the sample. This research consists of three steps, namely standardization, data clustering using K-means and data interpolation using the K-means clustering algorithm and zoning of 7 variables, namely magnitude, depth, victim died, the victim didn’t die, public facilities were heavily damage, public facilities were slightly damage, and affected areas.