Text Sentiment Analysis of Hotel Online Reviews

Abstract

Taking the Intercontinental Shenzhen as an example, this paper presents text analysis of hotel online reviews, including sentiment analysis based on sentiment dictionary and LDA topic analysis. The professional sentiment dictionary formed by expansion experiment improves the accuracy of sentiment analysis of online hotel reviews, which can be effectively applied to various scenarios of hotel online review sentiment analysis. Furthermore, LDA topic analysis is made for positive and negative reviews, in which perplexity index is introduced to determine the appropriate number of topics. Moreover, the output of LDA model is visualized by PyLDAvis. Through LDA topic analysis, the factors that lead to different guest reviews can be identified, which can further make recommendations for hotel operation management.