Recommender system providing recommendations for unidentified users of a commerial website

Recommender systems are a popular trend in recent research in Internet technologies. The models and algorithms of these systems are based on applying information of users and website content as well as their interconnections. However, there is a problem of applying these models and algorithms when the users are unidentified and there is an information gap to give recommendations. Each study case appears as a pair of a situation and a set of possible recommendations with their characteristics. The paper offers to solve the problem of generating recommendations through the adaptation process of the initial set of recommendations into which the recommendations are included on basis of criteria of similarity of the main and recommended contents. The adaptation process uses the iteration formula optimizing the recommendation utility function.