Article Per Year (5 Year)
p-Index From 2020 - 2025
0.23
P-Index
This Author published in this journals
All Journal Journal Of Informatics And Busisnes
Prakoso, Indra
Unknown Affiliation
Computer Science & IT Economics, Econometrics & Finance

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Sentiment Analysis of TikTok User Reviews on Google Playstore Using Naïve Bayes Methods Prakoso, Indra; Andhika Aziz Bachtiar; Elkin Rilvani
Journal Of Informatics And Busisnes Vol. 3 No. 2 (2025): Juli - September
Publisher : CV. ITTC INDONESIA

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47233/jibs.v3i2.3297

Abstract

interactions, one of which is TikTok. The TikTok platform has become a global phenomenon favored by many, especially the younger generation. As the number of users increases, reviews on digital platforms such as the Google Play Store become an important source for understanding users' perceptions of the application. Therefore, a deep understanding of user sentiment toward TikTok is essential for better app development and effective marketing strategies. To analyze TikTok user sentiment, this study employs two well-established computational methods: Support Vector Machine (SVM) and Naïve Bayes. These methods are used to classify user reviews into positive or negative sentiment categories. The approach involves several stages, including data collection, data preprocessing, data splitting, sentiment classification, and model evaluation. The study shows that the SVM model achieved an accuracy of 88.76% with an AUC of 92.61%, outperforming Naïve Bayes, which achieved an accuracy of 84.27% and an AUC of 92.57%. In the positive sentiment category, SVM recorded a precision of 90.74% and a recall of 95.15%, while Naïve Bayes yielded a precision of 83.61% and an almost perfect recall of 99.03%. For negative sentiment, SVM showed a precision of 80.39% and recall of 67.21%, whereas Naïve Bayes had a higher precision of 91.30% but a lower recall of 34.43%, with a lower F1-score of 50%.
Co-Authors Andhika Aziz Bachtiar Elkin Rilvani