Transformer-based Generative Adversarial Network (GAN) hybrid model for sentiment analysis

Information Technology

Sentiment analysis, which focuses on extracting insights from textual data to reveal people's opinions or attitudes on a specific issue, has become a prominent area of research in natural language processing (NLP), especially with the rise of social media. Twitter, as a popular platform, allows users to openly share their views and thoughts. However, analyzing Twitter data presents unique challenges due to the frequent use of slang, abbreviations, and misspellings in its short-form content. Traditional automated feature selection methods face limitations, including increased computational costs as the number of features grows. Deep learning, with its ability to self-learn and efficiently process large datasets, helps address these challenges. This paper proposes using a conditional generative adversarial network (GAN) for sentiment analysis on Twitter, with a convolutional neural network (CNN) employed to extract features from the data. Compared to previous approaches, the proposed method achieves superior results in terms of accuracy, recall, precision, and F1 score, with a classification accuracy of 93.33%.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.