Analisis Pengklasteran Data Pajak terhadap Pendapatan Daerah dengan Metode K-Means dan K-Median di Kabupaten Langkat

Abstract

Tax is one of the main sources of income for local governments in financing various development programs and public services. Langkat Regency, which is located in North Sumatra Province, alsi relies on taxes as a component of its regional income. K-means is one of the most widely used non-hierarchical clustering methods, but because it uses the mean as the cluster center, this method is more sensitive to the presence of outliers in the data. The K-median method, which uses the median as the cluster center value, is considered to be able to overcome outliers. This study aims to classify and describe data on restaurant taxes, hotel taxes, and swallow’s nest taxes in Langkat Regency in 2022 using the K-means and K-median methods. The clustering results show that in the first cluster k-means method there are 17 districts, in the second cluster there are 5 districts, and in the third cluster there is 1 district. Whereas in the k-median method in the first cluster there are 1 sub-district, in the second cluster there are 2 sub-districts, and in cluster 3 there are 20 sub-districts.