Sentiment Analysis of Netizens' Comments About Dwi Sasetyaningsih on the TikTok Platform Using a Deep Learning-Based Method Long Short Term -Memory (LSTM)

Ditha Aditya Pernikasari (Universitas Padajaran)
Tristy Welas Ulas Jati (Padjadjaran University)

Proceedings of Jogjakarta Communication Conference (JCC)
Vol. 4 No. 1 (2026): The 7th Jogjakarta Communication Conference (JCC)

The digital age, marked by rapid advancements in information technology, has transformed social media from merely a source of entertainment into a space where the public expresses opinions on various public issues. One such issue that has sparked a wide range of reactions on social media is the statement made by Dwi Sasetyaningsih, a recipient of the LPDP scholarship, regarding her child’s citizenship. This study aims to analyze sentiment in text data by applying a deep learning method based on Long Short-Term Memory (LSTM), and drawing on Bing Liu’s theory of Sentiment Analysis and Opinion Mining. The data used consists of a collection of comments from TikTok platform, which were then categorized into two sentiment classes: positive and negative. The research results produced an LSTM model from a dataset of 3,395 comments with a best accuracy of 87.48%. “positive” comments accounted for 72.90%, while “negative” comments accounted for 27.20%. It was concluded that the TikTok platform differs from other social media platforms in sentiment analysis of comment sections using text analysis methods.