Research On Text Sentiment Analysis Based On Topic Model

With the development of the Internet era,more and more users like to shop on ecommerce platforms,at the same time,a lot of user comment data is also generated,because these comment data contain a large number of user's attitudes and user's information about the topic of interest.Therefore,mining these data can provide some reference information for other consumers and businesses.Based on this background,this article uses the JST model to perform sentiment analysis on the comment text data.In order to effectively obtain the features and sentiment topic information in the comment text data,this article uses feature parts of speech to process the text to reduce the sparseness of the text,and then uses the Word2 vec method and the external sentiment dictionary to obtain the text data prior knowledge.Then,the processed comment data and the sentiment annotation files of the text vocabulary are applied to the JST model to obtain the sentiment polarity of the vocabulary in the corpus and the top vocabulary for different topics under different sentiment polarity tags.Finally,the sentiment information obtained above is used to analyze sentiment tendency of product reviews.Finally,this article applies the product review data to the JST model,analyzes and compares the topic acquisition between the LDA topic model and the JST model,and the acquisition and classification prediction of the sentiment topic of the JST model under different text sentiment polarity annotation files.The experimental results show that the sentiment analysis using the sentiment polarity annotation text obtained by this method is more effective and accurate.