Implementation Of Naïve Bayes Algorithm For Sentiment Analysis Of Creator Content On Tiktok Social Media

Abstrak

Social media has become a crucial space for content creators to build interactions, build self-image, and generate economic opportunities. TikTok, as a rapidly growing platform, generates a variety of public opinions in the form of comments reflecting positive, negative, and neutral sentiments toward creators. Sentiment analysis using a text mining approach is considered effective for understanding audience opinion in real time. This study applies the Naïve Bayes algorithm to classify the sentiment of TikTok user comments. This method was chosen for its speed and efficiency in text analysis, although it has limitations due to the assumption of independence between features. Model performance was evaluated using accuracy, precision, recall, and F1-score metrics to assess classification performance. The results showed that the Naïve Bayes algorithm was able to classify comments with an accuracy of 84.39%. The precision obtained was 0.862, recall was 0.844, and the F1-score was balanced, proving this method's effectiveness for sentiment analysis on TikTok comments. These findings confirm the potential of Naïve Bayes as a fast and accurate classification method for understanding audience opinion on social media.