Paper Key : IRJ************957
Author: Dhruv D. Patel,Milan Gohel,Bhavesh J. Soliya
Date Published: 06 Apr 2025
Abstract
Understanding public sentiment through social media content is very important but is a challenging task due to the existence ofsarcasm and context-dependent expressions. By adding context awareness and sarcasmdetection, this study extends the accuracy of sentiment analysis. Qualitative content analysis was used in conjunction with quantitative sentiment scoringin a mixed-methods approach. They gathered six months worth of social media posts and then employed sophisticated natural language processing methods context-aware transformers. We saw a 15% improvement in the accuracy of the enhanced sentiment analysis modelas well as significant improvements in precision, recall, and F1 score over traditional models. These were confirmed using regression analysis, wherea positive correlation was found between contextual understanding and sentiment classification accuracy. The model sustained stable performances across sixmonths. Most of the findings supported previousresearch, but a few suggested that sarcasm can be interpreted correctly when the context is strong and consistent. Further development of contextual modeling would enhance these aspects, and exploring contexts in multiple languages and mediums would help to more comprehensively understand customersentiments.
DOI Requested