MDE-DRIVEN TOURISM RECOMMENDATION SYSTEM

Tourism recommendation systems have become increasingly important as travelers rely on digital platforms to discover destinations, plan itineraries, and evaluate travel experiences. Traditional recommendation systems often face several challenges, including limited personalization, coldstart problems, static recommendation rules, and lack of transparency in recommendation logic. To address these issues, this research proposes an MDE-Driven Tourism Recommendation System that integrates Model-Driven Engineering (MDE) with hybrid machine learning techniques to provide adaptive and explainable travel recommendations. The proposed framework introduces a JSON-based Domain Specific Language (DSL) that allows dynamic modification of recommendation policies, weighting parameters, and decision constraints without requiring changes to backend code or system redeployment. The system utilizes a hybrid recommendation approach that combines contentbased filtering using TF-IDF and cosine similarity with collaborative filtering through matrix factorization to generate personalized destination suggestions. Additionally, sentiment analysis is applied to user reviews to assess qualitative attributes such as safety, affordability, cleanliness, and overall visitor satisfaction. An automated training mechanism continuously updates the recommendation models based on new user feedback and interaction data, thereby improving system accuracy over time. The architecture is designed using a scalable full-stack framework that supports secure authentication, real-time data interaction, and modular extensibility. Furthermore, the system incorporates an explainable recommendation module that improves transparency by presenting the reasoning behind each suggested destination. Experimental evaluation demonstrates that the proposed system improves recommendation accuracy, adaptability, and user trust compared with traditional tourism recommendation models. The framework provides a flexible and intelligent solution suitable for both academic research and practical tourism platforms.