MACHINE LEARNING ANALYTICS FOR PREDICTING TAX REVENUE POTENTIAL: AN EMPIRICAL CASE OF ANONYMOUS TAX OFFICE

• Main Author: David Febriminanto, Raden
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/67765
• Institution: Institut Teknologi Bandung

In line with rapid business process digitalization in the tax directorate, the size of data stored in the institution has grown exponentially. The issue of how to generate value out of the valuable data assets remains unsolved. Correspondingly, this research provides machine learning-based predictive analytics as a solution on how to use the taxpayers’ trigger data as a decision support system to discover and realize unexplored tax potential. More specifically, this research presents predictive analytic models that can accurately predict which potential taxpayers are likely to pay their due. We develop three machine learning models, namely logistic regression, random forest, and decision tree. We analyzed a total of 5,562 tax revenue potential data samples with eight predictors such as trigger data nominal value, distance tax office, type of taxpayer, media of tax report, type of tax, type of report status, and registered year of the taxpayer, and area coverage. Our study shows that the random forest model provides the best prediction performance. The results of the weight of each attribute indicated that the status of the tax report is the top tier of variable importance in predicting tax revenue potential. The analytics can help tax officers in determining potential taxpayers with the highest likelihood of paying their due. Given the size of the data records, this approach can provide tax administrations with a powerful tool to increase work efficiency, combat tax evasion, and provide better customer service.