Application of the C45 Decision Tree Method in Evaluating The Potential and Contribution of Retribution to Pad: Case Study of Barru Regency

The effectiveness of local revenue (PAD) collection is often constrained by the inability of traditional evaluation methods to identify key factors affecting the contribution of levies. This study aims to evaluate the potential and contribution of levies to PAD using the C4.5 Decision Tree method, offering a data-driven alternative for more precise fiscal analysis. Employing a quantitative approach, the study analyzed twelve months of tax data from four sectors—Hotel, Parking, Entertainment, and Advertising Taxes—in Barru Regency. The C4.5 algorithm was implemented using Weka 3.8 and validated through 10-fold cross-validation, with performance measured by accuracy (91.7%), precision (0.89), recall (0.87), and F1-score (0.88). The results revealed that Entertainment and Hotel Taxes contributed most significantly during certain seasonal peaks, while Parking Tax showed stable contributions year-round. The derived decision tree model successfully identified key tax attributes that influence PAD categorization into low, medium, and high contributions. This analytical framework enables more nuanced policy formulation, such as adjusting entertainment tax rates during high seasons and incentivizing compliance among hotel operators. The study underscores the potential of algorithmic approaches to improve fiscal management in the public sector and provides a replicable model for other regional administrations aiming to optimize local tax revenues..