Text Sentiment Analysis Based on CNN-BiGRU Enhanced Features
Jiahui Zhu, Haiming Li · 2022
In order to solve the problem that traditional neural network models tend to ignore the importance between aspect words and context, and can not fully obtain and utilize the corresponding features in aspect sentiment analysis task.A neural network model of convolutional neural network(CNN) and bidirectional gated recurrent unit(BiGRU) based on attention mechanism to enhance the aspect words features is proposed.The model enhances feature information related to aspect words through CNN and BiGRU.Then The enhanced features are used by BiGRU to obtain global features, and finally the attention mechanism is used to screen out the most relevant features of the global features and get the emotion category of the text.Experimental results show that compared with other baseline models, this model can obtain more relevant aspect features.The accuracy and F1 values are better than other models.