Research on Deep Learning Algorithm for Sentiment Analysis of Business English Texts

Emotional information contained in business English text data is of great value to enterprise decision-making. In this study, a deep learning model integrating long short term memory (LSTM) and attention mechanism was constructed to improve the ability to capture key emotional information, thus improving the accuracy of emotional analysis. By introducing the pre-trained BERT model as a feature extractor, the model can make full use of the rich language knowledge and feature representation ability in the pre-trained model and improve its performance in new tasks. The experimental results show that the proposed model is superior to other benchmark models and experimental variants in accuracy, macro-average F1 score and weighted average F1 score, especially in dealing with complex emotions and long texts.