Sentiment analysis in hospitality using text mining: The case of a Portuguese eco-hotel

Abstract

The rapid development of the Internet and mobile devices enabled the emergence of travel and hospitality review sites, leading to a large number of customer opinion posts. While such comments may influence future demand of the targeted hotels, they can also be used by hotel managers for improving customer experience. Nevertheless, this trend poses a problem, considering information is widely scattered, making almost impossible to extract from it useful knowledge. In this study, with the aim of facilitating this process, sentiment classification of an eco-hotel is assessed through a text mining approach using several different sources of customer reviews. Two dictionaries are compiled for building the lexicon used to parse the 401 reviews collected from a Portuguese eco-hotel between January and August of 2015. Then, the latent Dirichlet allocation (LDA) modeling algorithm is applied to gather relevant topics that characterize a given hospitality issue by a sentiment. Findings of this study state that accuracy is influenced by interaction between LDA generated topic models and the correct construction of both dictionaries. These results also reveal that text mining can generate new insights into variables that have been extensively studied in hospitality industry, including that hotel food generates ordinary positive sentiments for the case studied, while hospitality generates both ordinary and strong positive feelings. Such results are valuable for hospitality management, validating the approach proposed.