Dependency Tree-based Sentiment Classification using CRFs with Hidden Variables

In this paper, we present a dependency tree-based method for sentiment classification of Japanese and English subjective sentences using conditional random fields with hidden variables. Subjective sentences often contain words which reverse the sentiment polarities of other words. Therefore, interactions between words need to be considered in sentiment classification, which is difficult to be handled with simple bag-of-words approaches, and the syntactic dependency structures of subjective sentences are exploited in our method. In the method, the sentiment polarity of each dependency subtree in a sentence, which is not observable in training data, is represented by a hidden variable. The polarity of the whole sentence is calculated in consideration of interactions between the hidden variables. Sum-product belief propagation is used for inference. Experimental results of sentiment classification for Japanese and English subjective sentences showed that the method performs better than other methods based on bag-of-features.