A Review on Hybrid Clustering and Machine Learning Approaches for Sentiment Analysis
K. M. Poornima, J. Joselin · International Journal of Versatile Research and Analysis · 2026
Sentiment analysis has become a fundamental task in Natural Language Processing (NLP), Artificial Intelligence (AI), and Data Mining due to the increasing volume of user-generated content on social media, e-commerce platforms, blogs, and online forums. Traditional supervised machine learning algorithms require large volumes of labelled data, whereas clustering techniques can discover hidden structures in unlabelled datasets. Hybrid approaches that combine clustering and machine learning leverage the strengths of both paradigms by improving data representation, reducing labelling effort, enhancing classification accuracy, and handling noisy data. This review examines the evolution of hybrid clustering and machine learning methods for sentiment analysis. It discusses preprocessing techniques, feature engineering, clustering algorithms, classification algorithms, hybrid frameworks, benchmark datasets, evaluation metrics, applications, challenges, and future research directions. The review also highlights the integration of transformer models and Generative AI with hybrid learning frameworks, identifying open research challenges and opportunities for developing scalable, explainable, and multilingual sentiment analysis systems.