Sentiment Analysis of X Users on Iconnet Service Provider Using Naïve Bayes and Support Vector Machine
Hani Handayani, Rice Novita, Inggih Permana, Megawati Megawati · 2024
This research responds to the rapid growth of users of Iconnet, a subsidiary of PLN, PT Indonesia Comnets Plus (ICON+), which offers reliable, affordable, and unlimited internet services. To improve its services, Iconnet needs a deeper understanding of user opinions about its connections and services. The main objective of this research is to fill this gap by performing sentiment analysis of Iconnet users' opinions on X. Using a dataset of 2720 data, this research applies the Naive Bayes Classifier (NBC) and Support Vector Machine (SVM) algorithms to classify user opinions into three classes: negative, neutral, and positive. The 10-fold Cross-Validation method is used to improve the validity of the results. Results show the superiority of the SVM model, with accuracies on connection sentiment of 92% (SVM) and 78% (NBC), and on service sentiment of 88% (SVM) and 85% (NBC). Thus, this research not only provides insight into user evaluations of Iconnet but also highlights the superiority of SVM in sentiment analysis of internet services.