PRIVACY-PRESERVING COLLABORATIVE FILTERING

摘要

The recommender system may generate a predicted item rating for one user by performing cooperative filtering on item ratings from other users. The recommender system may include a client device that interfaces with a server to obtain a predicted item rating for a local user. The client device can generate a standardized rating vector for the user and calculates the group identifier for the user based on the standardized rating vector. The system may also generate a noisy rating vector for the local user and send a user-rating snapshot containing the group identifier and the noisy rating vector to the recommendation server. The recommender system may also include a recommendation server that generates a predicted item rating for a user by performing cooperative filtering on rating vectors from a plurality of other users belonging to the same rating group.