Sentiment Analysis of English Movie Reviews using Deep Learning

As an essential research direction in Natural Language Processing (NLP), Sentiment analysis technology aims to identify and classify emotional tendencies in text data automatically. This study proposes a simplified version of a movie review sentiment analysis model. The model is designed to reduce dependence on computational resources, making it more suitable for edge computing environments. The study draws on the basic architecture and techniques of deep learning models from the image processing field to achieve this goal, creating a more straightforward yet still powerful structure. The model primarily comprises convolutional and fully connected layers, maintaining efficient data processing capability even in situations with limited computational resources. Experimental results indicate that, despite its simplified structure, the proposed sentiment analysis model performs exceptionally in processing large volumes of online review data. Specifically, the model achieves an accuracy rate of 87%, significantly reducing computational costs and providing a new pathway for implementing sentiment analysis quickly and effectively in environments with limited computational resources.