This study aims to analyze aspect-based sentiment regarding the performance of the TikTok app based on user reviews on Google Play, as well as to identify changes in user perceptions over time and across app versions following app updates. The performance aspects analyzed include speed, stability, and responsiveness. This study employs a quantitative method with a descriptive approach. Data was obtained through web scraping of TikTok user reviews on Google Play during the period of January 1–December 31, 2025. The data was processed through the stages of text preprocessing, sentiment labeling using the INSET lexicon, aspect labeling using a keyword-based approach, feature representation using TF-IDF, and classification using the Support Vector Machine (SVM) method. The results of the study indicate that SVM is capable of effectively classifying sentiment across all aspects of the app’s performance. Speed and stability were the most frequently discussed aspects and were dominated by negative sentiment. Additionally, the analysis revealed that app updates influenced changes in user perceptions of TikTok’s performance throughout 2025.Keywords: aspect-based sentiment analysis, TikTok, SVM.
Copyrights © 2026