Description of Invention
This project explores the use of advanced language models—BERT, RoBERTa, and their ensemble for analyzing consumer behavior based on the consumer review dataset. By training these models for sentiment classification, this project aim to uncover patterns in consumer feedback. To ensure model transparency and trust, SHAP is applied to interpret predictions and highlight key features influencing sentiment decisions. This study aims to enhance understanding of consumer sentiment while ensuring transparency and explainability in AI-driven decision-making through the integration of natural language processing and explainable AI techniques.