Research on Aspect-Level Emotion Categorization Based on Hybrid Attention Mechanism

Aspect-level sentiment classification is a fine-grained sentiment analysis task that aims to classify different aspects of sentiment in text. Aiming at the problems of shallow feature extraction and weak generalization ability in aspect-level sentiment classification, this paper proposes an aspect-level sentiment classification model based on the fusion hybrid attention mechanism. In this paper, we propose an aspect-level sentiment classification model based on the fusion hybrid attention mechanism. First, the aspect words, locations, and contextual information of the text are extracted from the dataset. Second, the aspect word, location, and context information are encoded by BERT. Finally, text features are extracted through multiple and local attention mechanisms, and the features are fused for learning. Experiments demonstrate that the results on the publicly available Restaurant, Laptop dataset and Twitter dataset show significant improvement in accuracy compared to the traditional baseline model.