EMOTION RECOGNITION IN LEARNERS WITH EMOJI SENTIMENT ACCOMPANIMENT USING THE PHOBERT MODEL
DOI:
https://doi.org/10.18173/2354-1059.2024-0034Keywords:
opinion mining, sentiment analysis, emotion recognition, Emoji, BERT, PhoBERTAbstract
This paper proposes an advanced method for recognizing learners' emotions by incorporating the use of emojis to reflect the modern communication tendencies of learners, typically young individuals. The method is built on the PhoBERT model, a variant of BERT optimized for Vietnamese. Data was collected from opinion surveys of learners at the Ho Chi Minh City campus of the University of Transport and Communications to train and test the model. The system is designed to analyze text and recognize seven basic emotions: enjoyment, trust, hope, sadness, surprise, fear, and others. Corresponding emojis are then assigned to each emotion type to more clearly illustrate the learners' emotional states. Experimental results show that combining PhoBERT and emojis not only enhances the accuracy of emotion recognition but also makes communication more intuitive and vivid. The model achieved an accuracy of 74.1%. The paper also discusses practical applications of this system in the field of education, where teachers can quickly and accurately understand and respond to students' emotions, thereby improving teaching effectiveness.
References
[1] Devlin J, Chang MW, Lee K & Toutanova K, (2019). BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies.
[2] Liu Y, Ott M, Goyal N, Du J, Joshi M, Chen D & Stoyanov V, (2019). RoB-ERTa: A Robustly Optimized BERT Pretraining Approach. arXiv preprint arXiv:1907.11692.
[3] Sun C, Huang L & Qiu X, (2019). Utilizing BERT for Aspect-Based Sentiment Analysis via Constructing Auxiliary Sentence. Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies.
[4] Yang Z, Dai Z, Yang Y, Carbonell J, Salakhutdinov R & Le QV, (2019). XLNet: Generalized Autoregressive Pretraining for Language Understanding. Advances in Neural Information Processing Systems, 33, 1-11.
[5] Nguyen DQ & Nguyen AT, (2020). PhoBERT: Pre-trained language models for Vietnamese. Proceedings of the 2020 Conference on Empirical Methods, Natural Language Processing: Findings.
[6] Felbo B, Mislove A, Søgaard A, Rahwan I & Lehmann S, (2017). Using millions of emoji occurrences to learn any-domain representations for detecting sentiment, emotion, and sarcasm. Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing.
[7] Barbieri F, Kruszewski G, Ronzano F & Saggion H, (2016). How Cosmopolitan Are Emojis?: Exploring Emojis Usage and Meaning over Different Languages with Distribution-al Semantics. Proceedings of the 2016 ACM on Multimedia Conference.
[8] Eisner B, Rocktäschel T, Augenstein I, Bosnjak M & Riedel S, (2016). emoji2vec: Learning Emoji Representations from their Description. Proceedings of The Fourth International Workshop on Natural Language Processing for Social Media.
[9] Pham QH, Nguyen VA, Doan LB, Tran NN & Thanh TM, 2020. From Universal Language Model to Downstream Task: Improving RoBERTa-Based Vietnamese Hate Speech Detection. 12th International Conference on Knowledge and Systems Engineering (KSE), 37-42.
[10] Navyasree G, Saka N, Narayana DRVS, 2019. Using Hashtags to Capture Fine Emotion Categories from Tweets. International Journal for Innovative Engineering & Management Research, 8(5), 288-296, SSRN.
[11] Ekman P, Ekman E & Lama D, 2018. The Ekmans’ Atlas of Emotion. Paul Ekman Group.
[12] Sau TNT & Toanh TQ, 2021. Application of Bert Architecture for Storage Time of Record Classification Problem. TNU Journal of Science and Technology, 267(7), 41-49, https://doi.org/10.34238/tnu-jst.3990.
