Automatic generation of rating requests from a recommender system

Abstract

The recommender system uses stored representations of relations between users. Access to an item by one or more first users in the storage system is recorded. When the recommender system detects a request for a rating of the item from a second user, the recommender system selects at least one of the first users that occurs in a stored relation with the second user. A query message that refers to the second user is then automatically transmitted to enter a rating for the item to a user address associated with the selected first users. Preferably at least one first user is selected for which it has been recorded that it has accessed the item but for which it has not yet given a rating. Preferably, the rating from the first user is passed to the second user via the recommender system, which records the rating for further use. The selection of the first user may be based on recorded data that measures correlation between ratings from the first and second user.

Dependencies are inferred from the claim text.