Sentiment analysis of Chinese stock reviews based on BERT model

A large number of stock reviews are available on the Internet. Sentiment analysis of stock reviews has strong significance in research on the financial market. Due to the lack of a large amount of labeled data, it is difficult to improve the accuracy of Chinese stock sentiment classification using traditional methods. To address this challenge, in this paper, a novel sentiment analysis model for Chinese stock reviews based on BERT is proposed. This model relies on a pre-trained model to improve the accuracy of classification. The model use a BERT pre-training language model to perform representation of stock reviews on the sentence level, and subsequently feed the obtained feature vector into the classifier layer for classification. In the experiments, we demonstrate that our method has higher precision, recall, and F1 than TextCNN, TextRNN, Att-BLSTM and TextCRNN. Our model can obtain the best results which are indicated to be effective in Chinese stock review sentiment analysis. Meanwhile, Our model has powerful generalization capacity and can perform sentiment analysis in many fields.