News-text Sentiment Classification Research Based on JST Model
Yan ZHAN, Hao CHEN, Hongyan MA, Guochun ZHANG
College of Mathematics and Information Science, Hebei University, Baoding, CHINA
Abstract: JST model is a hybrid model with both topic and sentiment, which adds emotional elements on the basis of the LDA model, and it can make text sentiment analysis and subject extraction in textual level. JST’s sentiment classification accuracy of the model is a little low, therefore we put forward the appraise dictionary for JST model of prior knowledge, which is to change the method assigning emotional tags randomly into the method assigning emotional tags the dictionary after comparing.
Keywords: Sentiment Analysis; JST Model; Appraise Dictionar