English Text Sentiment Analysis Network based on CNN and U-Net

Abstract

Sentiment orientation analysis of English text is a core issue within the realm of natural language processing. Traditional methods of word segmentation often encounter ambiguity when processing English language texts. In light of this, the present study introduces an innovative approach to English text sentiment analysis that utilizes a convolutional neural network (CNN) coupled with a U-network (U-Net). This method employs parallel convolutional layers to grasp the relationships and interactions between word vectors, which are then fed into a hierarchical attention network based on the U-Net to ascertain the sentiment polarity. The experimental outcomes demonstrate that the model achieves an accuracy of 93.45% in bias classification on an English review dataset, outperforming many existing sentiment analysis models.