Efficient Sentiment Analysis on IMDb Movie Reviews with Synonym Augmentation and Global Average Pooling

With the advancement in the field of automated recommendations, efficient sentiment analysis of reviews has become immensely important. This study presents the application of the Global Average Pooling (GAP) model in efficient sentiment analysis that has traditionally been dominated by models such as Recurrent Neural Networks. Focusing on IMDb movie reviews this paper showcases the effectiveness of the GAP model in handling text data in extracting features and reducing dimensionality. Through this approach, we observed improvements in classification accuracy while using an efficient model. Our findings suggest that the GAP model could serve as an alternative to conventional sentiment analysis techniques. Furthermore, our research indicates benefits from exploring the integration of GAP, with deep learning methods to further enhance performance.