• Sakshi Kumari
• Charu Kavadia
• Sumayya Fatima
• Tahlil Ullah Shaikh
• Yashashav Goyal

Sentiment analysis (SA) has been a crucial area of study within natural language processing (NLP), progressing from early rule-based systems to more recent methods that use transformer models, deep learning (DL), and multimodal techniques. This work offers a systematic evaluation of research contributions from 2005 to 2025, emphasising important techniques such sentiment-specific embeddings, hybrid models, lexicon-based approaches, classical machine learning, and new multimodal architectures. Particular attention is paid to the integration of symbolic thinking and audio-visual elements. The paper provides practical applications, performance benchmarks, and outstanding issues such as sarcasm detection and domain adaptability by combining these various approaches. Transformer-based models and multimodal integration have been the main topics of recent contributions [1, 2, 4], whereas earlier studies established the framework using symbolic techniques that are lexicon-based and symbolic [13, 16].

Deep Learning, Transformer Models, Lexicon-Based Models, Natural Language Processing, Sentiment Analysis, Multimodal Sentiment Analysis, Hybrid Techniques, and Emotion Recognition

"The Evolution of Sentiment Analysis in NLP:A Review of Techniques and Applications", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 6, page no.250-253, June-2025, Available :http://www.jetir.org/papers/JETIRGW06039.pdf

"The Evolution of Sentiment Analysis in NLP:A Review of Techniques and Applications", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 6, page no. pp250-253, June-2025, Available at : http://www.jetir.org/papers/JETIRGW06039.pdf