SASE: Sentiment Analysis with Aspect Specific Evaluation Using Deep Learning with Hybrid Contextual Embedding

In recent years, sentiment analysis has grown more intricate as the need for deeper insights from text data has expanded. Traditional methods fall short for capturing subtle opinions, giving rise to aspect-oriented sentiment analysis. This study proposes a new framework called Sentiment Analysis with Aspect-Specific Evaluation (SASE) fusing with diverse word embeddings to give aspect-specific sentiment analysis. This novel hybrid approach holds the promise of unravelling multifaceted sentiment aspects across varied domains, and when coupled with the robust RoBERTa model, demonstrates good improvements in accuracy with 78%. The comparison study of the SASE framework with baseline models are also discussed in this work.