ABSTRACT
Digitalization is supporting interdisciplinary studies combining computers and statistics, which is further enhanced by availability of abundant data being provided by firms. This has popularized the application of machine learning approaches in various disciplines to provide a statistical interpretation of models. Machine learning (ML) is a collection of algorithm that helps applications to predict the outcomes accurately without unambiguously programming it. It is based on the premise of developing algorithms that receive an input data and use statistics to forecast an output while apprising outputs as a fresh dataset becomes accessible. Advances in internet technology have boomed the online service sector in recent times. With the service sector gaining momentum over the years, researchers feel that the relationship between the service providers and customers must be further explored, especially under an online framework. Online researchers define customer satisfaction as an optimistic feeling that arises in reaction to the aggregate experience during his or her browsing sessions. Customers perceive the level of satisfaction on the basis of vital features of hotel services, and so these attributes play a major role on satisfaction behavior. It is also believed that a satisfied customer will be willing to spread word or recommend that service to his family and friends as well as try to make a repurchase. This buzz has been termed as electronic word-of-mouth (EWOM) or user-generated content (UGC). The emergence of digital medium has popularized the use of UGCs over a firm’s platform, which is easily accessible to potential purchasers. Visitors have opportunities to express their experiences and provide suggestions to others for hotels by making use of feedbacks on staff, parking, and cleanliness. This chapter aims to evaluate the experience of guests towards a hotel and the attributes that motivate in achieving the satisfaction level. Few studies have attempted to discuss the hotel guest satisfaction with the help of online reviews commented by them. Here, we propose a unique study that combines text mining technique along with machine learning approach to determine the satisfaction status of hotel guests and its determinants with the help of overall ratings, which can be considered as a proxy variable for guest satisfaction. The chapter aims to evaluate guest satisfaction from 2,39,859 hotel reviews extracted from Tripadvisor. First, topic modeling is applied with the help of latent semantic analysis (LSA), which results in topics that represent the whole useful review information regarding dimensions resulting in satisfied behavior. Then, for measuring satisfaction level, we make use of classifiers, such as naïve Bayes (NB), decision tree (DT), random forest (RF), support vector machine (SVM), and artificial neural networks (ANN), to check their accuracy using various performance measures. LSA findings show that the key dimensions are nightlife, value, amenities, natural beauty, guest experience (critique), recommendation, staff, guest experience (praise), the location of hotel, accessibility, car parking, visitor suitability (critique),satisficing, style and décor, bathroom, deals, visitor suitability (praise), hybrid, room experience, and high standards. Also, the RF classifier provides the best performance in comparison with others. Lastly, the studyprovides few theoretical and managerial inferences that might be helpful for a firm dealing with hospitality sector.
