ANALISIS SENTIMEN ULASAN PENGGUNA APLIKASI WEBTOON MENGGUNAKAN TEXT MINING DAN ALGORITMA SVM

LINE Webtoon is a popular application for reading comics online, available on the Google Play Store. It has been around since 2015 and has a large user community. Out of a total of 35 million active LINE Webtoon users worldwide, approximately 6 million of them are from Indonesia. Despite the app's popularity and many 5-star reviews, not all users are satisfied with the performance of LINE Webtoon. This can be seen from user reviews in the comment section on the Play Store. New users often use these reviews as a reference to determine the best and most satisfying app to use. In this research, a sentiment analysis of LINE Webtoon users was conducted with the aim of classifying user reviews into positive and negative sentiments. The classification algorithm used was Support Vector Machine (SVM). The study collected a total of 15,000 data sets through web scraping from 2018 to 2023. The data was divided into three testing groups, with 90% for training and 10% for testing, 80% for training and 20% for testing, and the last one with 70% for training and 30% for testing. The results of testing with 90% training data and 10% testing data showed a precision score of 0.86, a recall score of 0.86, an f- measure score of 0.86, and an accuracy score of 0.82. The results of testing with 80% training data and 20% testing data showed a precision score of 0.86, a recall score of 0.85, an f-measure score of 0.85, and an accuracy score of 0.81. Lastly, the results of testing with 70% training data and 30% testing data showed a precision score of 0.85, a recall score of 0.85, an f-measure score of 0.85, and an accuracy score of 0.81. After analyzing the results of all three testing scenarios, the highest accuracy was achieved in the testing with 90% training data and 10% testing data, which was 0.82 or 82%.

Keywords : Comment, Sentiment, Support Vector Machine, Webtoon