Sentiment Analysis on Twitter Data Using Acrgruatt: Amalgam Convolution Reversible Gated Recurrent Unit with Attention Network

Twitter (currently known as X), a popular social media network, allows users to share their ideas and opinions in real time using short, concise tweets. Despite the effectiveness of deep learning and natural language processing in emotion detection, the length of tweets poses hurdles for correctly classifying sentiment, resulting in decreased efficiency and accuracy. In response, this study introduces an Amalgam Convolution Reversible Gated Recurrent Unit with Attention Network (ACRGRUATT) for sentiment analysis (SA), which mainly includes three stages: preprocessing, feature extraction, and prediction. This approach uses an innovative amalgam feature acquisition technique to acquire optimal syntactic and semantic information, improving SA precision and efficiency. In addition, the authors propose a unique sentiment prediction architecture, the CRGRUATT. Using this technique, the approach mitigates the shortcomings of existing frameworks and advances SA on Twitter data. The findings demonstrate that the proposed strategy performs best with an accuracy of 97.84 per cent, a recall of 98.20 per cent, a precision of 97.52 per cent, and an F1 of 98.20 per cent with a minimal loss of 0.0445. By incorporating these techniques, the authors aim to enhance SA using Twitter data.