EMOTEX-PR: a multimodal transformer-enhanced framework for real-time product review sentiment analysis across diverse datasets

Key Points

• The framework processes diverse datasets for accurate sentiment analysis, demonstrating efficiency.
• A notable improvement of 25% in sentiment prediction accuracy was observed across multiple datasets.
• Analysis utilized a multimodal transformer approach to integrate various types of data for better performance.
• Further validation is necessary on real-world applications and broader datasets to confirm generalizability.