From Physical Agents to Recommender Agents

In this paper, we argue in favour of the interchange of research results between physical agents and recommender systems. Certain properties, such as introspection and physical foundations of physical agents, are mapped into agents that model and make recommendations to users from the subjective, content-based point of view of products and services. How this real-to-virtual mapping of properties and behaviours can be performed is the main goal of our work, from which we expect to obtain a general concept of recommender agents which improves the state of the art in terms of performance and maintainability. Furthermore, new features of the recommender agents could be also mapped to the physical agents discipline. Some of the features explained are installed in two personalization company products: Proto-agent and Habitat-Pro.