A Hybrid Fuzzy-Neural Framework for Multimodal Sentiment Classification
DOI:
https://doi.org/10.63282/3050-922X.ICAILLMBA-118Keywords:
Hybrid Neuro-Fuzzy Architecture, Sentiment Analysis, Mamdani Fuzzy Inference System, Emoji Semantics, Sarcasm Detection, Natural Language Processing (Nlp), Bi-LstmAbstract
The swift advancement of digital communication has created a need for more effective techniques for sentiment analysis and classification. Conventional Natural Language Processing (NLP) models frequently encounter difficulties with the subjective nature and linguistic subtleties present in social media. This research paper introduces an innovative hybrid architecture that combines deep learning with Fuzzy Logic to analyze a multi-dimensional set of features, incorporating 27 important emoji characteristics. The proposed approach utilizes a dual-stream pipeline: a Bidirectional LSTM (Bi-LSTM) captures deep semantic structures from the text, whereas a Mamdani-type Fuzzy Inference System (FIS) addresses the "gray areas" of human emotion. By associating emoji intensity with triangular membership functions μ(x), the model proficiently "defuzzifies" sarcasm and ambiguous meanings. Experimental findings reveal that this neuro-fuzzy methodology achieves an accuracy of 89.5% and an F1-score of 0.88, marking a considerable enhancement over baseline text-only techniques. In particular, the model performed well in clarifying sarcastic ambiguities, the situations where the text's polar meaning and the emoji's intent are at odds that are boosting ambiguity resolution.
References
[1] L. A. Zadeh, "Fuzzy sets," Information and Control, vol. 8, no. 3, pp. 338–353, 1965.
[2] B. Pang and L. Lee, "Opinion mining and sentiment analysis," Foundations and Trends in Information Retrieval, vol. 2, no. 1–2, pp. 1–135, 2008.
[3] H. Felbo, A. Mislove, A. Søgaard, I. Rahwan, and S. Lehmann, "Using millions of emoji occurrences to learn any-domain representations for detecting sentiment, emotion and sarcasm," in Proc. Conf. on Empirical Methods in Natural Language Processing (EMNLP), 2017, pp. 1615–1625.
[4] T. J. Ross, Fuzzy logic with engineering applications, 4th ed. Hoboken, NJ, USA: Wiley, 2016.
[5] S. Hochreiter and J. Schmidhuber, "Long short-term memory," Neural Computation, vol. 9, no. 8, pp. 1735–1780, 1997.
[6] P. Kralj Novak, J. Smailović, B. Sluban, and I. Mozetič, "Sentiment of emojis," PLOS ONE, vol. 10, no. 12, p. e0144296, 2015.
[7] P. Thakur, "A neuro-fuzzy framework for sentiment analysis on noisy social media data," Journal of Artificial Intelligence, Machine Learning and Soft Computing, 2025.
[8] K. Dave et al., "SentiQNF: A novel approach to sentiment analysis using quantum algorithms and neuro-fuzzy systems," IEEE Trans. Comput. Soc. Syst., 2025.
[9] A. Khan et al., "Sentiment analysis of emoji fused reviews using machine learning and BERT," Scientific Reports, vol. 15, 2025.
[10] Y. Lou et al., "Sentiment classification of emoji and text: An analysis of model families, fusion strategies, and performance gains," Communications on Applied Nonlinear Analysis, 2024.
[11] A. Ben Meriem et al., "A fuzzy approach for sarcasm detection in social networks," Procedia Computer Science, vol. 192, pp. 2400–2409, 2021.
[12] M. Hasan et al., "Sarcasm detection over social media platforms using hybrid ensemble model with fuzzy logic," Electronics, vol. 12, no. 1, 2023.
[13] K. P. Gowda et al., "Transformers in sentiment analysis: A paradigm shift in deep learning research," Journal of Information Systems Engineering and Management, vol. 10, 2025.
[14] A. Zadeh et al., "Multimodal language analysis in the wild: CMU-MOSEI dataset and interpretable dynamic fusion," in Proc. Assoc. Comput. Linguistics (ACL), 2018, pp. 2236–2246.
[15] J. S. R. Jang, "ANFIS: adaptive-network-based fuzzy inference system," IEEE Transactions on Systems, Man, and Cybernetics, vol. 23, no. 3, pp. 665–685, 1993.
[16] C. T. Lin and C. S. G. Lee, Neural Fuzzy Systems: A Neuro-Fuzzy Synergism to Intelligent Systems, Upper Saddle River, NJ, USA: Prentice Hall, 1996.
[17] Z. Huang, W. Xu, and K. Yu, "Bidirectional LSTM-CRF models for sequence tagging," arXiv preprint arXiv:1508.01991, 2015.
[18] J. Pennington, R. Socher, and C. D. Manning, "GloVe: Global vectors for word representation," in Proc. Conf. on Empirical Methods in Natural Language Processing (EMNLP), 2014, pp. 1532–1543.
[19] X. Hu et al., "We know how you feel: Sentiment analysis of emojis," in Proc. 11th Int. Conf. on Web and Social Media (ICWSM), 2017, pp. 112–121.
[20] D. Maynard and M. A. Greenwood, "Who cares about Sarcastic Tweets? Investigating the Impact of Sarcasm on Sentiment Analysis," in Proc. 9th Int. Conf. on Language Resources and Evaluation (LREC), 2014, pp. 4238–4243.
[21] R. Confalonieri et al., "Using Fuzzy Logic to Leverage Explainability in Social Media Sentiment Analysis," IEEE Transactions on Fuzzy Systems, vol. 29, no. 12, pp. 3624–3635, 2021.
[22] A. Adadi and M. Berrada, "Peeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI)," IEEE Access, vol. 6, pp. 52138–52160, 2018.
