Analisis Sentimen terhadap Fenomena Artis Berjualan di Tiktok Shop pada Media Sosial X Menggunakan Model Naïve Bayes

Abstract

TikTok is a popular application used for entertainment and income generation. One of its key features, TikTok Shop, offers opportunities for entrepreneurs to promote their products. However, issues arise as the public must compete with celebrities who also sell on the platform. To understand public responses, a sentiment analysis was conducted on tweets from the social media platform X (Twitter) discussing celebrities selling on TikTok Shop. From 1,189 collected tweets, text pre-processing was performed, resulting in 770 clean tweets. The data was then labeled, split, and analyzed using the Naïve Bayes Classifier method in Google Colab with Python programming. Manual labeling by language experts showed higher accuracy compared to automatic labeling using TextBlob. The results showed 494 neutral tweets, 253 positive tweets, and 23 negative tweets, with the highest accuracy of 88.31% achieved by the Multinomial Naïve Bayes method. In conclusion, the majority of the public holds a neutral stance toward the phenomenon of celebrities selling on TikTok Shop