Naive Bayes-Based Classification of Priority Service Applications at the Ministry of Religious Affairs of North Tapanuli

Abstract

This study aims to develop a classification model for determining service priority levels in administrative processes at the Ministry of Religious Affairs of North Tapanuli by examining the effectiveness of the Naïve Bayes algorithm in processing categorical service data. A simulated dataset of 200 service request records reflecting common administrative characteristics was created using a quantitative technique. Preprocessing, feature encoding, and train-test splitting were then performed before modeling with the Multinomial Naïve Bayes classifier. Accuracy, precision, recall, and f1-score were used to assess the model's performance; confusion matrix analysis and visualization were also used. The findings demonstrate that Naïve Bayes attained an accuracy of 78%, which is in line with earlier research showing the algorithm's competitive performance on categorical and public service datasets. These results suggest that administrative data with simple feature distributions can benefit from probabilistic models like Naïve Bayes. The paper indicates that Naïve Bayes can be used as a decision-support tool for administrative service prioritization and suggests further research to improve forecast reliability using actual operational data and algorithm comparison.