Speech Emotion Analysis Based on Bidirectional Long Short-Term Memory

Speech emotion analysis, as a crucial direction in the field of natural language processing, holds great significance in multiple fields such as business services, education, and academic research. It can help enterprises gain insights into consumers’ needs, assist education in tailoring teaching methods to different students, and promote academic research while enhancing the intelligent interaction experience. This study focuses on the emotion classification task of models based on the Bidirectional Long Short-Term Memory (Bi-LSTM) network for processing an open speech dataset. A self-attention module is incorporated, and comparative experiments and ablation experiments are designed. The results show that the Bi-LSTM model has achieved good classification results in the classification of this dataset.