Khushi Gautam and Rupali Bhartiya

Systematic Review of the Architectural Transition in Sentiment Analysis: From Hybrid Deep Learning to Generative Large Language Models

The findings show that although discriminative fine-tuned models continue to perform excellently in narrow classification tasks, generative LLMs possess impressive zero shot reasoning and flexibility but have issues with inconsistency and lack of transparency.