An unsupervised approach for sentiment classification

In this paper we propose a new unsupervised and domain independent approach for sentiment classification. It takes a few documents as the training set to build a sentiment vocabulary list which will be used to classify the documents according to their sentiment orientation. The system is self-supervised and domain independent. Experimental results show that the classification accuracy of the approach can reach 85.7% which is better than the previous experiments of unsupervised methods.