Abstract

Sentiment analysis is a domain within Natural Language Processing (NLP) focused on the computer detection and classification of opinions in textual data. The expansion of social media platforms, such as Twitter, Facebook, and Instagram, has resulted in an increase in user-generated content that reflects public opinion on various issues, including films, products, and political events, daily. An examination of this information would benefit firms, lawmakers, and other stakeholders by aiding in the evaluation of public perception. The traditional method of sentiment analysis relies on a rule-based framework and fundamental machine learning algorithms, including Naive Bayes, Support Vector Machines (SVM), and Logistic Regression (LR). These solutions generally necessitate manually generated features and face challenges in capturing more profound linkages and dependencies within the text. Recent advancements in deep learning, particularly in Recurrent Neural Networks (RNN) and Long Short-Term Memory (LSTM) networks, have transformed the domain of sentiment analysis by offering models for the unsupervised learning of associations from raw text data. As a variation of RNN, LSTM addresses the vanishing gradient problem, making it more suitable for tasks requiring longer dependencies, such as sentiment analysis. This study investigates the application of LSTM networks for binary sentiment categorisation utilising the IMDb movie review dataset. The model's learning capabilities were enhanced by using pre-trained word embeddings (GloVe), which illustrate semantic relationships among words and augment the model's contextual comprehension.

References

• Baccianella, S., Esuli, A., & Sebastiani, F. (2010). SentiWordNet 3.0: An augmented lexical resource for sentiment analysis and opinion mining. Proceedings of LREC 2010.
• Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8), 1735-1780.
• Kim, Y. (2014). Convolutional Neural Networks for Sentence Classification. Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), 1746-1751.
• Maas, A. L., Daly, R. E., Pham, P. T., Huang, D. S., & Ng, A. Y. (2011). Learning word vectors for sentiment analysis. In Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies (ACL HLT 2011), pp. 142-150.
• Mikolov, T., Sutskever, I., Chen, K., Corrado, G., & Dean, J. (2013). Distributed representations of words and phrases and their compositionality. Proceedings of the 26th Annual Conference on Neural Information Processing Systems (NeurIPS 2013), 3111-3119.
• Pennington, J., Socher, R., & Manning, C. D. (2014). GloVe: Global vectors for word representation. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP 2014), pp. 1532-1543.
• Varadharajan, V., et al. (2025). Deep Learning-Based Sentiment Analysis: Enhancing IMDb Review Classification with LSTM Models. ResearchGate. https://www.researchgate.net/publication/388184987_Deep_Learning-Based_Sentiment_Analysis_Enhancing_IMDb_Review_Classification_with_LSTM_Models
• Alasmari, M., Farooqi, N. A., & Alotaibi, Y. (2024). Sentiment analysis of pilgrims using a CNN-LSTM deep learning approach. PMC (PubMed Central). https://pmc.ncbi.nlm.nih.gov/articles/PMC11784790/