PREDICTION OF PUBLIC SERVICE SATISFACTION USING C4.5 AND NAÏVE BAYES ALGORITHM

Abstract

One of the things that has often been questioned lately is in the field of public services, especially in terms of the quality or service quality of government agencies to the community, the Manpower and Transmigration Office of Kab. Karawang is a government agency in charge of public services. where one of the tasks is to make an AK.1 card (yellow card), based on this problem the Manpower and Transmigration Office of Kab. Karawang Regency. Karawang seeks to improve service quality in order to satisfy consumers by distributing questionnaires to every consumer who is making an AK card.1. In this study, we will apply the C4.5 and Naïve Bayes algorithms to predict the satisfaction of public services with the nominal type of dataset used. The evaluation is done based on a comparison of the level of accuracy, precision, recall, and F-Measure using a confusion matrix. From the research that has been carried out, the Naïve Bayes algorithm with 70% training data distribution and 30% testing is able to provide better predictive results than the C4.5 algorithm as evidenced by the accuracy value = 96.89%, precision = 95.50%, recall = 95.00% and f-measure = 94.60%.