감성분석을 활용한 레스토랑 선택속성과 만족도에 관한 연구: 외래관광객 온라인 리뷰 중심으로

Online reviews contain various emotions of customers and can be used as a resource to measure and improve customer service experiences. Traditionally, survey methods are used to measure customer satisfaction, and research using text mining techniques also serves as a means to supplement the traditional method. This study is to verify that the sentiment polarity value of selection attributes has a significant effect on overall customer satisfaction by using online restaurant reviews left by tourists. A web crawler was created using open source Python to collect review and data was collected from the hotel & restaurant review community. After classifying review sentences by food, price, service, and atmosphere based on topic modeling, sentiment analysis techniques were used to calculate sentiment polarity values for each restaurant selection attribute. Also, Tobit regression model was used to verify the effect of selection attributes on customer satisfaction. As a result, it was found that food, price, and service variables had a significant effect on customer satisfaction, whereas atmosphere variable had relatively no significant effect compared to other variables. This result is meaningful in that the selection attribute, which has been considered important in the study of the survey method for tourists, was verified as an important factor in the study using the text mining method as well. The implications of this study are that text review information quantifies customer experience and measure customer satisfaction, and the research method through text mining can bring meaningful results along with the traditional approach of measuring customer experience through existing surveys.