<![CDATA[Currently, customer satisfaction is a crucial indicator in evaluating the quality of Fiber to the Home (FTTH) services. This study aims to analyze customer satisfaction with Indosat HiFi services and compare the performance of the Random Forest and Gaussian Naive Bayes algorithms in predicting customer satisfaction levels. The research dataset consists of Quality of Service (QoS) parameters and customer operational data, including latency, throughput, packet loss, downtime, response time, and complaint count. The data was processed using a Machine Learning approach through preprocessing, model training, and performance evaluation stages. The research results showed that Random Forest produced the best performance with 98.8% accuracy, 100% recall, 99.4% F1-score, and 95.9% CV Mean. Meanwhile, Gaussian Naïve Bayes obtained 97.6% accuracy, 98.8% F1-score and 95.2% CV Mean. The research findings show that service quality and user experience factors influence the level of customer satisfaction. The resulting model has great potential to support data-based decision making to improve the quality of FTTH services. The research data was processed and then tested with a ratio of 80:20. Model evaluation was carried out using accuracy, recall, F1-score and cross validation.]]>
Copyrights © 2026
