Hybrid IndoBERT-CNN Model for Sentiment Analysis: A Case Study on Educational Public Figures on TikTok

Indar Muslimah Kusmadi

doi:10.36914/jrtk.v1.i1.11

Authors

Indar Muslimah Kusmadi Universitas Tarakanita Author

DOI:

https://doi.org/10.36914/jrtk.v1.i1.11

Keywords:

Sentiment Analysis, IndoBERT, CNN, Deep Learning, Tik Tok, University Branding

Abstract

This study public sentiment toward educational content shared by the Rector of President University on TikTok, positioning the platform as a both a learning medium and a tool for institutional branding. A dataset of more than 2,000 user comments was collected through web scrapping and preprocessed to retain Indonesian language texts. Sentiments were automatically labelled into positive, neutral, and negative categories, and then classified using a hybrid deep learning architecture that integrates IndoBERT for contextual embeddings with Convolutional Neural Network(CNN) for enhanced feature extraction. The proposed model was trained and validated on the dataset, achieving strong accuracy and reliable performance compared to baseline machine learning approaches. The analysis revealed that the majority of comments expressed positive sentiment, indicating favorable audience engagement with the Rector’s educational content, while neutral and negative comments highlighted areas for improvement in communication style and message framing. These findings demonstrate the potential TikTok as an effective channel for academic leaders to disseminate educational message, enhance institutional visibility, and strengthen public trust in higher education. Beyond technical contributions, this research provides actionable insight for universities, strategies in digital promotions and student recruitment, highlighting the societal impact of educational engagement via social media.

References

Devlin, J., Chang, M. W., Lee, K., & Toutanova, K. (2019). BERT: Pre-training of deep bidirectional transformers for language understanding. Proceedings of NAACL-HLT 2019, 4171–4186. Association for Computational Linguistics.

Dhendra, & Gayuh Utomo, V. (2025). Benchmarking IndoBERT and transformer models for sentiment classification on Indonesian e-government service reviews. Jurnal Transformatika, 23(1), 86–95.

Jayadianti, R., et al. (2024). Sentiment analysis of Indonesian reviews using fine-tuning IndoBERT and R-CNN. ILKOM Jurnal Ilmiah, 14(3), 348–354.

Kim, Y. (2014). Convolutional neural networks for sentence classification. Proceedings of EMNLP 2014, 1746–1751. Association for Computational Linguistics.

Koto, F., Rahimi, A., Lau, J. H., & Baldwin, T. (2020). IndoLEM and IndoBERT: A benchmark dataset and pre-trained language model for Indonesian NLP. Proceedings of COLING 2020, 757–770.

Mandhasiya, D. G., Murfi, H., & Bustamam, A. (2024). The hybrid of BERT and deep learning models for Indonesian sentiment analysis. Indonesian Journal of Electrical Engineering and Computer Science, 33(1), 591–602.

Murfi, H., Syamsyuriani, T., Gowandi, T., Ardaneswari, G., & Nurrohmah, S. (2022). BERT-based combination of convolutional and recurrent neural network for Indonesian sentiment analysis. arXiv preprint arXiv:2202.09812.

Riskia, A. S., Wufron, & Roji, F. F. (2025). Analisis sentimen Coretax: Perbandingan pelabelan data manual, transformers-based, dan lexicon-based pada performa IndoBERT. MALCOM: Indonesian Journal of Machine Learning and Computer Science, 5(3), 1037–1048.

Setiawan, J. C., Lhaksmana, K. M., & Bunyamin, B. (2022). Sentiment analysis of Indonesian TikTok review using LSTM and IndoBERTweet algorithm. JIPI (Jurnal Ilmiah Penelitian dan Pembelajaran Informatika), 8(3).

Wu, Y., Jin, Z., Shi, C., Liang, P., & Zhan, T. (2024). Research on the application of deep learning-based BERT model in sentiment analysis. arXiv preprint arXiv:2401.12345.