부산대학교 도서관

This study provides a framework to discover the strengths and weaknesses of services and to enhance customer satisfaction from online customer reviews, especially in the hospitality industry. Instead of using user-rating scores, a sentiment analysis is conducted to classify consumers’ attitudes toward hotels based on their experiences as written in reviews. By integrating the results of sentiment analysis, polarity, with a structural topic model as a covariate, this research detects hidden dimensions in review data and distinguishes positive and negative dimensions. Furthermore, other covariates, time and hotel, are used to track dimensional trends over time and compare the outcomes of two different hotels. This study addresses a limitation in previous research that focused on discovering the service defect by strength detection and explanation of the relationship between service factors and user-rating scores of the review as a proxy of customer satisfaction.