Travel time context-based recommendation system using content-based filtering / Nurin Syazwani Amran and Salehah Hamzah

Tourism, defined as travelling to various locations for pleasure, has long been crucial to a country's economic development. However, the rapid expansion of the tourism industry, particularly in Europe, has brought about challenges linked to information overload for travellers seeking suitable destinations and optimal travel times. These challenges manifest in two significant technological aspects. Firstly, the vast amount of tourism-related information available on the internet poses a daunting task for individuals to identify suitable travel destinations. Information from various sources, such as websites, blogs, and newspapers, needs more organization, making it overwhelming for visitors. This often leads travellers to make misguided choices, causing dissatisfaction if their selected destination aligns differently from their preferences. Secondly, tourist itinerary planning faces challenges in obtaining precise information about the optimal time context to visit diverse destinations. Travellers frequently rely on guidebooks, online platforms, or recommendation systems, which require optimization for factors like time feasibility. This complexity increases the likelihood of travellers missing experiences best enjoyed at particular times. In response to these challenges, this research introduces a personalized travel recommendation system for Malaysian tourists exploring select European countries—namely, the United Kingdom, Germany, France, Switzerland, and Italy. Employing content-based filtering techniques, the system considers user profiles and preferred travel times to provide tailored recommendations, enhancing the overall travel experience. The prototype, implemented as a user- friendly web application, aims to offer comprehensive guidance on optimal travel times and destinations. Functionality testing indicates a successful implementation, albeit with minor interface and data handling issues. For future recommendations, the prototype will incorporate advanced machine learning techniques, such as exploring hybrid models that integrate collaborative filtering, to enhance recommendation accuracy and overall performance.