Classification of Corporate Tax Compliance in Indonesia Based on k-Nearest Neighbors Algorithm

This study demonstrates the application of machine learning for classifying the tax compliance level of manufacturing companies in Indonesia. The k-nearest neighbor (kNN) algorithm was utilized to develop the machine learning model. A dataset was collected through a survey conducted directly with finance personnel in charge of the manufacturing companies. Data collection proved challenging as not all companies were willing to participate, resulting in 209 data points. Accountants analyzed the collected data and identified three classes of tax compliance levels. The developed machine learning model successfully achieved a classification accuracy of 92.86%.