Analisis Sebaran Titik Rawan Bencana dengan K-Means Clustering dalam Penanganan Bencana

Abstract

Puwakata Regency has fertile land, agricultural products and abundant natural resources. However, the area is also vulnerable to disaster risk. Based on the data collected, the disasters that occurred in Puwakata Regency included several categories, namely landslides, droughts, hurricanes and floods. The trend of increasing numbers of disasters requires further investigation to prevent an increase in the number of victims. Given the large amount of data available, this information can be obtained through data mining analysis methods. For natural disaster data, the clustering method in data mining is very useful for grouping disaster data based on the same characteristics, so that it can be used as a basis for classifying future disaster events. The k-means algorithm is a model used to form clusters by measuring how close it is to the data set. Therefore, in terms of the location of the disaster, the type of disaster and its impact on the disaster, it is hoped that this research can use the clustering technique with the k-means algorithm to classify disaster-prone points. The results obtained 3 clusters, namely, the type of drought disaster is cluster 0, the type of landslide is cluster 1, and the type of landslide is cluster 2. After forming three clusters, disaster management strategies are drawn up at each disaster-prone point in the Purwakarta area