A Content-Based Smart Tourist Recommendation System Using TF-IDF and Cosine Similarity

Tourism recommendation systems have become essential tools for assisting travelers in making informed decisions from a large volume of available travel information. Traditional tourism platforms often provide generic suggestions that fail to adapt to individual preferences. This paper presents a Smart Tourist Recommendation System based on content-based filtering techniques using TF-IDF vectorization and cosine similarity. The proposed system analyzes attraction descriptions, categories, and related metadata to generate personalized recommendations without relying on historical user data. A multi-page web interface is implemented using Streamlit to support attraction-wise, city-wise, and comparison-based recommendations. Experimental evaluation shows that the system efficiently delivers relevant recommendations with low computational complexity, making it suitable for real-world smart tourism applications.