Volume : VII, Issue : XI, November - 2018

Unsupervised Sentiment Analysis on Chinese Microblog Based on Topic Sentiment Model

Runyu Liang, Lanhe Yang

Abstract :

 Sentiment analysis in microblog has important theoretical and application value in personalized recommendation and public opinion analysis.In order to solve the problem of feature sparsity of microblog corpus, Then the paper uses the weight of the words in corpus as one of the BTSM‘s (Biterm Topic–Sentiment Model) parameter to form the unsupervised W–BTSM model finally. The model adds a sentiment layer to Biterm Topic Model and fuses the weighted model with it ,thus a three–layer Bayesian model of "sentiment–topic–term" is formed. It directly models the generation process of biterm in microblog corpus. The experimental results on the NLP & CC2012 corpus and the real micro–blog data obtained through web crawler show that W–BTSM model can effectively identify the sentiment tendency of Chinese microblog, the F–measure of W–BTSM model is higher than ASUM model and JST model. < clear="all" style="page–eak–before:always;mso–eak–type:section–eak" />

Keywords :

Article: Download PDF   DOI : 10.36106/ijsr  

Cite This Article:

Unsupervised Sentiment Analysis on Chinese Microblog Based on Topic Sentiment Model , Runyu Liang, Lanhe Yang , INTERNATIONAL JOURNAL OF SCIENTIFIC RESEARCH : Volume-7|Issue-11| November-2018

Number of Downloads : 255

References :