Machine Learning-Based Sentiment Analysis in English Literature: Using Deep Learning Models to Analyze Emotional and Thematic Content in Texts

This paper proposes a hybrid deep learning approach combining Bidirectional Long Short-Term Memory (BiLSTM) networks and an attention mechanism to extract sentiment and thematic content from literary texts. The model is designed to capture complex emotional nuances and themes in literature by processing text data from both forward and backward directions, while the attention mechanism enables the model to focus on the most important sections of the text. Hyperparameter optimization is performed using the Improved Particle Swarm Optimization (IPSO) algorithm to fine-tune the model for efficient sentiment extraction. A case study using a dataset of 500 English novels spanning various genres demonstrated the effectiveness of the proposed approach. The model achieved high accuracy and F1 scores in sentiment classification and thematic extraction, outperforming traditional methods like CNNs. The analysis revealed key emotional themes such as joy, fear, and sorrow, and the thematic content included love, betrayal, and revenge. The results highlight the potential of deep learning to advance literary analysis by providing deeper insights into both emotional and thematic layers of literary works. Future directions include exploring multimodal data integration and expanding the application of deep learning in the humanities.