SRNN-MAFM-based Unimodal Text Sentiment Analysis

Sentiment analysis, a crucial task in natural language processing (NLP), involves extracting subjective information from text data to understand public opinion, market trends, and brand reputation. This research introduces a groundbreaking unimodal text analysis system called Stacked-Recurrent-Neural-Network with Multi-level Attention and Feedback module (SRNN-MAFM) for sentiment analysis. The proposed system employs advanced techniques to perceive sentiment in texts and categorizes them into three extreme sentiments: positive, neutral, and negative. The sentiment analysis module utilizes a Boolean structure and an OR operation for classification. The study includes a Python platform simulation using the STS-Gold datasets to evaluate the proposed technique. The simulation results demonstrate that the proposed unimodal text sentiment analysis technique achieves an impressive accuracy of 97.8%, outperforming existing methods. This research contributes to the advancement of sentiment analysis and provides a highly accurate and efficient solution for analyzing sentiment in unimodal text data.