Klasifikasi Pemukiman Kumuh di Wilayah DKI Jakarta Menggunakan Decision Tree

Abstract

Jakarta as the capital of Indonesia has a high level of urbanization, causing a high rate of population density. High population growth in a small area tends to create a slum in the area. So far, the Statistics Central Bureau (BPS) has conducted a survey of the slum areas, but the method used is still quite subjective. To reduce the problem of subjectivity, this study aims to apply Decision Tree in classifying the slum areas in Jakarta and determine the accuracy of the classification process. The data used in this study consisted of 10 parameters, namely, population density, building layout, construction of houses, residential ventilation, building density, state of roads, drainage/sewerage, water consumption, human waste disposal, and waste management. There were 320 sample data collected using stratified random sampling. The data were divided into 2 groups: training data and testing data. Training data were used to construct a Decision Tree, while testing data were used to test the tree. Testing process used 10-fold cross validation resulting in an accuracy of 69.7%, 68.8%, and 67.4% for confidence factors of 0.1, 0.2, and 0.3, respectively.