ANALYZING TAX DEBT PAYMENT BEHAVIOR USING DECISION TREE ALGORITHM

Active tax collection is crucial for securing state revenue, particularly when economic performance falls short of fiscal targets. Although the Directorate General of Taxes has intensified enforcement efforts such as issuing warning letters, distress warrants, asset confiscation, and account freezing, their effectiveness remains limited. This study examines the interplay between active tax collection, tax burdens, and taxpayer non-compliance in shaping tax debt payment behavior. Employing a decision tree algorithm, the analysis is based on taxpayer data from the Bandung Cicadas tax office. The results indicate that active tax collection is dominant in encouraging payment, with distress warrants being the most influential factor, particularly for taxpayers with lower tax burdens. In contrast, higher tax burdens and persistent non-compliance reduce the likelihood of repayment. The study also highlights how psychological and financial considerations influence payment behavior, particularly under moderate tax burdens. These insights support the relevance of Behavioral Economics in understanding taxpayer responses to enforcement measures.