Penerapan Algoritma K-Nearest Neighbor Untuk Klasifikasi Rumah Layak Atau Tidak Layak Huni (Studi Kasus: Desa Bulu Kecamatan Kraksaan Kabupaten Probolinggo)

Abstract

The house is a need that must be met, in addition to food and clothing needs, as well as an indicator of community welfare, to create a safe, comfortable and healthy living environment. The government provides social rehabilitation assistance to the community for uninhabitable houses by helping to buy building materials so they can rehabilitate homes that are not proper. In Bulu Village there are still many places to live that are categorized as uninhabitable houses. This is because of the community's income factor and limited knowledge about the function of the house, as well as facilities and infrastructure that make it increasingly difficult to realize livable dwellings. This research uses the K-Nearest Neighbor algorithm for the classification of decent or uninhabitable homes which aims to provide convenience in determining prospective recipients of social rehabilitation assistance for uninhabitable houses with accurate results and minimizing mistargeting. The results of this study are an average accuracy of 96.25% with a standard deviation of 5.73% in the 7th model test using 10-fold cross-validation with odd k and validation and evaluation of results with the confusion matrix getting an accuracy value of 100%, the precision value is 100%, the recall value is 100%, and the f-measure value is 100% with k = 13.