Micro-blog has become increasingly popular among the general public. It has brought a lot of comment text to researchers for its great convenience, timely updating, and a wide variety of self-focused topics. Identifying the emotions expressed in these comments has become a valuable topic in order to make inferences for focused contents in Micro-blog. In this paper, we report on the effectiveness of the language representation model BERT [1] with respect to the sentiment classification tasks. Experimental results show that the pre-training of deep bidirectional transformers can improve the accuracy, recall and F1 score on sentiment classification. The final evaluation index of this problem by using a Github data set increased by 2.3% on average.
CITATION STYLE
Zheng, J., Chen, X., Du, Y., Li, X., & Zhang, J. (2020). Short Text Sentiment Analysis of Micro-blog Based on BERT. In Lecture Notes in Electrical Engineering (Vol. 590, pp. 390–396). Springer Verlag. https://doi.org/10.1007/978-981-32-9244-4_56
Mendeley helps you to discover research relevant for your work.