Tourism Recommender System using Machine Learning Based on User's Public Instagram Photos

In the past decade, recommender systems have become an essential part of online services such as NetFlix, YouTube, online shopping, etc. The tourism agencies such as TripAdvisor or Expedia also apply the recommender system to their services. For Thailand, the tourism industry is one of the most important revenues of the country. The problem is that the recommender system for planning a trip to Thailand still not effective enough. Users require a lot of effort when planning a trip. Therefore, the objective of this study is to develop the prototype of a tourism recommender system that automatically understands the user's preferences of their favorite tourist attractions without asking them any question. It applied machine learning to extract the user's preferences from the user's Instagram photos. Those preferences then use to compute the similarity with the attributes from 23 example tourist attractions in Ubon Ratchathani Province. A user study was conducted with 42 participates to preliminary study the precision and the adoption of the prototype. The results suggested that the prototype has been judged as satisfactory by participants for both precision and adoption. Moreover, the findings of this study will serve as insights for the direction of our planned future research such as applying the recommender system to other provinces of Thailand.