กรอบแนวคิดสำหรับระบบแนะนำโดยอาศัยบริบทของนักท่องเที่ยวเชิงกีฬา

Abstract

Currently, tourists' preferences for travel have become increasingly specific and targeted. For example, sports tourism has gained popularity, and tourists have a desire to explore destinations where they can enjoy while participating in related activities. However, existing recommendation systems are unable to fully accommodate the specific travel preferences of these tourists. Therefore, contextaware technology needs to be utilized in recommendation systems for sports-oriented tourists to enhance their travel experience. Sport tourism is a specialized form of travel where sports enthusiasts travel with the objective of participating in sports-related events. When searching for information on running competitions, competitors prefer to access the internet. However, the information on the internet is scattered, making this to be difficult for them to find the specific information they need. Therefore, the use of semantic web technology is necessary to provide structured and easily searchable information. This research presents the framework of a recommendation system that provides a list of running events aligning with the user's context. To recommend a specific running event, the framework incorporates two types of ontologies, running event and user profile ontologies. The structure and semantic of the developed ontologies have been improved and modified based on expert evaluations, with precision at 0.91 and recall at 0.98. Furthermore, the K-modes clustering technique for data classification and the Apriori algorithm for association rule discovery were used to develop a rule-based recommendation. The Jena Inference Rules were applied to convert the discovered association rules, enabling the recommendation of running events that align with the user's preferences. The research prototype is validated and evaluated with scenario information collected from 5 runners. The actual results of the prototype's recommendations were compared with the expected results designed from runner scenarios. The evaluation result has shown that the developed prototype is able to accurately and precisely recommend interested running events for sport tourists.