Sentiment Analysis of YouTube Comments on the Closure of TikTok Shop Using Naïve Bayes and Decision Tree Method Comparison

Abstract

As technology advances, YouTube has become a social media platform that allows users to watch, broadcast, and share videos. One of the videos that has garnered a lot of comments from the public is about the closure of TikTok Shop. This research uses two methods: Decision Tree and Naïve Bayes. The aim of this study is to compare the Naïve Bayes and Decision Tree methods in analyzing public sentiment regarding the closure of TikTok Shop. The test results for both methods are not significantly different. Each method is divided into three research scenarios. In Scenario 1, with an 80:20 data split, the Decision Tree method achieved an accuracy of 74.71%, a precision of 57%, a recall of 57%, and an F1-score of 57%, while Naïve Bayes had an accuracy of 73.96%, a precision of 58%, a recall of 34%, and an F1-score of 29%. In Scenario 2, with a 70:30 data split, the Decision Tree method achieved an accuracy of 73.27%, while Naïve Bayes achieved an accuracy of 73.99%. In Scenario 3, with a 60:40 data split, the Decision Tree method achieved an accuracy of 71.78%, while Naïve Bayes achieved an accuracy of 74.02%. The evaluation results indicate that the Decision Tree method using an 80:20 data split has superior accuracy compared to the Naïve Bayes method.