Odjemalski priporočevalni sistem na osnovi HATEOAS

This final project paper examines whether personalized recommendations can be generated without sending users' personal interaction data to centralized servers.To investigate this, a client-side recommender system based on a lightweight HATEOAS-inspired architecture was designed and implemented.The system infers user preferences from implicit feedback, specifically clicks and viewing duration, and generates recommendations locally using content-based filtering over movie genres.Because interaction data remains on the user's device, the approach supports privacy-preserving personalization.The system was evaluated in an online user study with 31 participants.The results showed that recommendation quality was rated significantly above the neutral midpoint, suggesting that meaningful personalization can be achieved while preserving user privacy.Kudiaeva I. Towards a client-side HATEOAS-based recommender system.