earticle

영어

Sentiment analysis, also referred to as opinion mining, has received increasing attention as a method for evaluating customer satisfaction with hotel services. By identifying the emotions behind a review, sentiment analysis can provide useful information to support quality management and marketing. With an aim of addressing the lack of sentiment-analysis models for hotel services in Korea, this paper presents a sentiment-analysis model for analyzing online reviews written in Korean. The presented model conducts a term-frequency analysis to figure out the most frequently mentioned service attributes and identifies whether the sentiment towards each attribute is positive or negative. The model also provides a list of words associated with each attribute, such that one can clarify why customers have such sentiments for the attribute. To demonstrate the model, this paper analyzes a total of 4,203 Korean reviews on 18 5-star hotels in Seoul, Korea. ‘Hotel A’ with the most number of reviews is taken as a representative case, and its detailed results are discussed. The results show that the sentiment analysis can serve as an effective means of understanding customer needs and assessing the degree of customer satisfaction with the hotel service. The strengths and weaknesses of the hotel service can be identified as well.