LAUSR

The development of social media has changed the way that travelers visit sightseeing spots. The social Internet of Things (IoT) allows products to automatically generate posts, share content and location information, and help build an online community of users based on their company’s products, so that marketing personnel can also get useful feedback and understand the user’s opinions. In tourism and hospitality industry, to enhance the revisit intention of passengers is an important issue for the purpose of increasing margin. In recent years, related researches had focused on the customers’ revisit behaviors and factors. However, few studies have investigated the related issues that travelers do not want to visit again. Failure to revisit may bring a great damage to the company’s revenue in the future. To avoid this situation, a text mining based approach will be proposed to identify non-revisit factors from online textual reviews in social media. Because it is impossible to determine whether a passenger has intention to revisit, this study proposed a text mining based approach which uses sentiment of text reviews to identify the passenger’s motivations (negative for revisit and non-negative for revisit). Then, feature selection methods, decision tree (DT), Least Absolute Shrinkage and Selection Operator (LASSO), and Support Vector Machines Recursive Feature Elimination (SVM-RFE) will be utilized to discover the important factors of non-revisit factor set. Back-propagation Neural Networks (BPN) and Support Vector Machines (SVM) will be employed to evaluate the effectiveness of selected feature sets. Finally, experimental results could be provided to travel service providers to improve service quality and effectively avoid non-revisit behaviors in the future.