Intent Classification from Online Forums for Phuket Medical Tourism

Abstract

Social media makes healthcare and medical information readily available to medical tourists. The medical tourists use social media for searching and communicating about their intents. As the questions posted on social media are rapidly increased, the difficulty to read all questions by human is increased as well. Hospitals running medical tourism business also need to know the needs of medical tourists for improving services and providing the right products to them. The needs or intents of medical tourists can be found on questions that they ask. Therefore, the objective of this study is to collect and classify intents of medical tourists from the questions posted on online forums. In this study, we collect questions related to medical tourism from the TripAdvisor website. We use natural language processing (NLP) to pre-process the questions and classify them using two neural network models, i.e., a BiLSTM model and a BERT model. The experimental result shows that the BERT model provides better performance with 94.22% of accuracy. We also analyze the results and summarize shortcomings of the dataset and the models.