Penerapan Algoritma Naive Bayes untuk Mengklasifikasikan Tingkat Kepuasan Pelanggan pada Fadhli Komputer

Fadhli Computer is a computer maintenance service business that requires structured evaluation to measure customer satisfaction levels in order to maintain loyalty and service quality. This study aims to classify customer satisfaction levels using the Naïve Bayes algorithm based on Service Quality dimensions with a 1–5 Likert scale. Primary data were collected through questionnaires from 140 respondents, which were then divided into 112 training data and 28 test data with three class labels: Very Satisfied, Satisfied, and Dissatisfied. The classification calculation steps were carried out manually and tested using RapidMiner software. The test results show that the Naïve Bayes algorithm successfully classified the distribution of test data consistently, detailing 23 very satisfied customers, 4 satisfied customers, and 1 dissatisfied customer. The evaluation of model performance using the confusion matrix resulted in a perfect accuracy, precision, recall, and F1-score of 100%. This study concludes that the Naïve Bayes algorithm is highly effective and reliable for classifying customer satisfaction at Fadhli Komputer, and can also serve as a basis for management decision-making.