Exploring Topic and Sentiment of Hotel Reviews

Abstract

Online review can help people getting more information about products. It offers a unique proposition for sentiment analysis. In this paper, a method for automatic extracting topic and it’s sentiment of hotel reviews is proposed. Our method focus on a special case which means some explicit topics is already given. The unknown topics are extracted by means of given topics. Previous studies mainly focus on the polarity of the sentiment. Furthermore, in this paper, the sentiment strength is calculated by lexicon-based approach. Empirical experiments on hotel reviews demonstrate the advantage of the proposed method.