Analisis Sentimen Ulasan Produk Layanan Internet di Twitter Menggunakan Naive Bayes Classifier

Abstract

This research aimed to conduct sentiment analysis on user reviews of Indihome, one of Indonesia's leading internet service providers, using data collected from Twitter. With a growing user base of 7 million, Indihome achieved an impressive revenue growth rate of 48% by the end of 2019. To assess user satisfaction, sentiment analysis was performed on tweets containing the keyword "indihome" between July 1 and July 11, 2022. A total of 10,000 tweets and retweets were collected and subjected to data preprocessing, manual labeling, and training using the Naive Bayes Classifier method. The training phase involved 1,000 tweets, equally divided between positive and negative sentiments, manually labeled for accuracy. For evaluation, 200 labeled tweets were used as testing data. The sentiment analysis results were remarkable, with 98% accuracy achieved for the training data and 91% for the testing data. These findings demonstrate the effectiveness of the Naive Bayes Classifier in accurately determining sentiment from Indonesian-language tweets about Indihome's services. By analyzing user opinions expressed on Twitter, this study provides valuable insights into the level of user satisfaction with Indihome as an internet service provider. Leveraging sentiment analysis applications on Twitter social media enables companies like Indihome to gain a deeper understanding of customer sentiment and make informed decisions to improve their services. The research showcases the importance and effectiveness of sentiment analysis in evaluating user feedback and enhancing customer satisfaction in the internet service industry.