Sentiment analysis is the computational study of emotions, opinions, and attitudes expressed in text. Over the years, this field has advanced from basic lexicon-based approaches, which relied on predefined word lists, to sophisticated machine learning and deep learning techniques capable of detecting and interpreting sentiment patterns across vast datasets. This paper provides a comprehensive, concept-centric literature review of sentiment analysis, examining its evolution, diverse models, and extensive applications. It critically addresses key challenges in the field, such as handling language complexities—including sarcasm, context dependency, and the necessity for cross-domain adaptability. Notably, this review introduces a novel framework for understanding sentiment analysis advancements and identifies emerging trends and research gaps. By evaluating the evolution and current state of sentiment analysis research, this paper not only highlights significant trends but also proposes new directions for future exploration. Ultimately, this work aims to deepen the understanding of this rapidly evolving domain and provide valuable insights for future research.
