Studi Tentang Algoritma C5.0 Dalam Memprediksi Kepatuhan Nasabah Dalam Membayar Pajak Pertambahan Nilai

Abstract

Data mining is the process of extracting valuable patterns, information, and knowledge from large data sets. Data mining has an important role in identifying and minimizing risks in various lives. One of the algorithms of data mining is the decision tree type C5.0. The C5.0 algorithm is an algorithm used to solve classification problems. The C5.0 method can be applied in various sectors such as the taxation sector. Paying taxes is an obligation by an individual or entity paid to the State. The value added tax is the highest contributory tax because it is collected several times to companies. Factors that affect customer compliance in paying value added tax are income, entity form, and reporting status. This study aims to predict public compliance in paying value added tax using the C5.0 method. This research aims to produce a classification model that can be a potential solution for dealing with prediction problems in customer compliance with paying taxes. The results of the study obtained an average accuracy of the C5.0 model of 66,5%. Based on this accuracy value, the model can be categorized as still weak in predicting the status of value added tax payments.