<![CDATA[TikTok is a popular application among young people. TikTok was an application initially launched in China before landing in Indonesia at the end of 2017. Unfortunately, the popularity of TikTok stems from personal lack of self-image, for example wearing sexy clothes, dancing in erotic and inappropriate moves. This is based on many positive and negative comments from TikTok users. So we need a way to automatically classify reviews through sentiment analysis. The purpose of this study is to classify TikTok user comments on Google Play Store using Naive Bayes and C4.5 algorithms. This study used 1330 data, of which 602 data were negative and 728 data were positive. The results show that the Naive Bayes algorithm produces accuracy values of 79.00%, 79.00% precision, 78.00% recall, and 78.00% F1 score. The C4.5 algorithm produces 68.00% accuracy, 68.00% precision, 68.00% recall, and 68.00% F1 score. We can conclude that the Naive Bayes algorithm is the best algorithm compared to the C4.5 algorithm. The Naive Bayes algorithm achieves an accuracy value of 79.00%.]]>
Copyrights © 2024
