Sentiment Analysis of #KaburAjaDulu Tweets Using the IndoBERT Deep Learning Model

• NURUL ZAHRAH
• 2025

---

This study aims to analyze public sentiment toward the #KaburAjaDulu hashtag on Twitter using IndoBERT, a deep learning model specifically designed for the Indonesian language. The hashtag
reflects public dissatisfaction and concern over Indonesia’s socio-economic and political conditions, especially among the younger generation. Tweets were collected through scraping, preprocessed,
and labeled into sentiment categories (positive, negative, neutral). After training and evaluation, the IndoBERT model achieved an accuracy of 96.75%, along with high and balanced precision, recall,
and F1-scores across all sentiment classes. The results demonstrate that IndoBERT is capable of accurately and consistently classifying sentiment in Indonesian-language tweets. This highlights the
potential of NLP-based approaches to automatically and effectively understand public opinion
expressed on social media.

• Sentiment Analysis
• IndoBERT
• Twitter
• #KaburAjaDulu
• Natural Language Processing
• Deep Learning

https://repository.bsi.ac.id/repo/67118/Sentiment-Analysis-of-#KaburAjaDulu-Tweets-Using-the-IndoBERT-Deep-Learning-Model
Anda dapat membaca file tugas akhir ini di repository.