Sentiment Analysis of Chinese Comment Text Based on CNN-BiGRU-ATTENTION

Sentiment analysis for Chinese comment data is a hot research topic today, and has significant implications for various fields of social life such as business, government, and education. This article selects comment data from 10 categories of products, integrates attention models, and proposes the CNN-BiGRU-ATTENTION model for sentiment analysis of Chinese comment data. Extract text features through CNN and BiGRU layers, and then introduce the extracted features into the attention mechanism to complete feature extraction of text information. Finally, obtain classification results through the Softmax layer. The experimental results show that the proposed model has improved accuracy compared to CNN, Text-CNN, GRU, BiLSTM, and CNN-ATTENTION models.