Real-Time Context-Aware Recommendation System for Tourism.

Area of Science:

• Computer Science
• Artificial Intelligence
• Tourism Informatics

Background:

• The tourism industry is evolving towards 'Tourism 2.0', emphasizing enhanced travel experiences and online information exchange.
• Current tourism recommendation systems struggle with insufficient data and real-time changes, limiting their effectiveness.
• Intelligent tourism service tools are needed for personalized recommendations, time savings, and marketing optimization.

Purpose of the Study:

• To propose a real-time recommendation system for tourism (R2Tour) that provides customized destination suggestions.
• To develop a system capable of responding to dynamic situations, including external factors and geographical information.
• To address the limitations of existing systems in handling insufficient or rapidly changing information.

Main Methods:

• Developed R2Tour, a machine learning-based system utilizing situational data (temperature, precipitation) and tourist profiles (gender, age).
• Trained models to recommend the top five nearby tourist destinations in real-time.
• Evaluated R2Tour's performance using six machine learning models (K-NN, SVM) and data from Jeju Island attractions.

Main Results:

• R2Tour achieved a recommendation accuracy of 77.3%.
• The system demonstrated strong performance with micro-F1 score of 0.773 and macro-F1 score of 0.415.
• The model effectively learned tourism patterns from situational information for real-time recommendations.

Conclusions:

• R2Tour successfully provides real-time, customized tourist destination recommendations by integrating situational awareness and user profiles.
• The system's ability to adapt to changing conditions offers a significant improvement over traditional recommendation engines.
• Future applications include in-vehicle recommendation systems and targeted advertising within the tourism sector.