Machine Learning Framework for Adaptive Museum Tour Recommendation and Personalized Visitor Experience

This study proposes a modified random forest (MRF) framework for designing adaptive museum tour systems that deliver personalized visitor experiences. Using the Rijksmuseum visitor behavior dataset, the system models visitor preferences according to dwell times, movement sequences, and exhibit interaction patterns. The MRF approach introduces temporal feature weighting and path continuity penalties to enhance prediction accuracy and logical tour progression. Experimental results demonstrate that MRF achieves a classification accuracy of 98.7% and a mean path coherence score (PCS) of 0.96 and that it significantly increases simulated visitor engagement in comparison to baseline methods. Comparative analysis with state-of-the-art models confirms the superiority of the proposed framework. This research advances adaptive cultural heritage technologies and offers a scalable, real-time solution for enriching museum experiences through intelligent, behavior-driven personalization.