Abstract
User-Generated-Content (UGC) has gained increasing attention as an important indicator of business success in the tourism and hospitality sectors. Previous literature has analyzed travelers' satisfaction through quantitative approaches using questionnaire surveys. Another direction of research has explored the dimensions of satisfaction based on online customers' reviews using the machine learning approach. This study aims to present a new method that combines machine learning and survey-based approaches for customers' satisfaction analysis during the COVID-19 outbreak. In addition, we investigate the moderating role of service quality on the relationship between hotels' performance criteria and customers' satisfaction. To achieve this, the Latent Dirichlet Allocation (LDA) was used for textual data analysis, k-means was used for data segmentation, dimensionality reduction approach was used for the imputation of the missing values, and fuzzy rule-based was used for the prediction of satisfaction level. Following that, a survey-based approach was used to validate the research model by distributing the questionnaire and analyzing the collected data using the Structural Equation Modeling technique. The result of this research presents important contributions from the methodological and practical perspectives in the context of customers' satisfaction in tourism and hospitality during the COVID-19 outbreak. The outcomes of this research confirm the significant influence of the quality of services during the COVID-19 crisis on the relationship between hotel services and travellers' satisfaction.