Abstract
Mining medical tourists' preferences and detecting their satisfaction level through Electronic Word of Mouth (eWOM) in medical tourism websites is an important task. Machine learning techniques have been very successful in developing recommendation agents through the analysis of eWOM in the e-commerce context. However, such methods are fairly unexplored in the medical tourism context through the analysis of user-generated content. This research is the first attempt to analyze eWOM in medical tourism websites for tourists' preferences mining using machine learning techniques. The results of the eWOM analysis revealed that the learning techniques are able to effectively analyze online reviews and accurately predict their preferences for their decision-making process in medical tourism. Compared to the methods which rely solely on the supervised learning techniques, the method evaluation results demonstrated that the use of fuzzy clustering and text mining approaches can be an important stage of eWOM analysis in the prediction of medical tourists' preferences.