Klasterisasi Tingkat Kemiskinan di Indonesia menggunakan Algoritma K-Means

Abstract

Poverty is one of the deep social challenges around the world and is a major focus in the global development agenda. This article discusses the role of clustering methods in analyzing and understanding poverty issues. We use data from Statistics Indonesia (BPS) on 34 provinces in Indonesia to classify groups of people who are vulnerable to poverty. Clustering analysis helps us identify characteristics that may be overlooked by conventional approaches, which in turn enables the development of more targeted and effective solutions to poverty. We use the K-Means method in our analysis and present it within the framework of the CRISP-DM methodology. The results show that almost 95% of the poor in Indonesia belong to the 'Poor' group. Therefore, we recommend effective actions based on indicators that are the main factors of poverty, as well as designing specific policies for regions with similar characteristics. This article aims to contribute to the global effort to end poverty and achieve the vision of equitable and inclusive sustainable development.