International Journal of Computer Techniques Volume 12 Issue 4 | Bi-LSTM Model for Emotion Recognition
Bi-LSTM Model for Emotion Recognition
International Journal of Computer Techniques – Volume 12 Issue 4, July – August 2025
ISSN: 2394-2231 | https://ijctjournal.org
Abstract
This study presents a Bidirectional Long Short-Term Memory (**Bi-LSTM**) deep learning model for **emotion recognition from text**. Designed to classify six core emotions—**Sadness, Anger, Love, Surprise, Fear, Joy**—the model utilizes forward and backward LSTMs to capture full contextual dependencies within sentence structure. Results show superior performance over baseline algorithms on key metrics including **accuracy, precision, recall**, and **F1-score**, offering practical use cases for mental health monitoring, social media analysis, and human-computer interaction.
Keywords
Deep Learning, LSTM, Bi-LSTM, Text, Emotion Recognition
Conclusion
The BiLSTM architecture proves effective for emotion classification in natural language text, outperforming traditional models by leveraging **bidirectional context** and sequential memory. Future work may integrate **attention mechanisms**, **transfer learning**, and larger corpora to improve precision and interpretability. This research contributes to the growing domain of emotion-aware intelligent systems.
References
- Andryushechkin et al. (2017). Workshop on Subjectivity & Sentiment Analysis. ACL.
- Ardiada et al. (2019). Text Mining on Social Media with SVM & KNN. Jurnal Informatika.
- Halim et al. (2020). RNN for Automatic Summarization. Jurnal Infra.
- Coursera: Tweet Emotion with TensorFlow. [Online Course]
- Riza & Charibaldi (2021). LSTM + FastText for Twitter Emotion. IJAIR.
- Rohman et al. (2020). Emotion Detection Using TF-IDF. CSRID.
Post Comment