Poverty classification using Analytic Hierarchy Process and k-means clustering

Abstract

The successfulness of poverty alleviation programs depends on the accuracy of poverty data. The government needs to collect poverty data and analyze them to determine which poverty alleviation programs should be delivered to. A data collection process is often done by conducting a survey that consists of 14 survey variables. However, raw data collected from surveys are not useful if they are presented as is. These survey data need to be processed further to support decision making. This paper presents a method to process survey data into categories using Analytic Hierarchy Process (AHP) and k-means clustering method. The categories consist of three poverty levels, such as near poor, poor, and very poor. We also present a workflow of survey and a implementation of this method to collect and process poverty data.