A personalized retrieval system with preserving privacy

The collection of extensive knowledge about users´ interests, behavior, and actions is necessary for most personalized retrieval systems. However, users´ browsing information and interest model contain their personal privacy, thus disclosure of privacy is possible. This paper proposes a personalized retrieval system called APIRS to preserve privacy. Its personalized service runs on the client side; the retrieval service runs on the server side; and the user login to the server is anonymous. The server only knows the client is a valid user, but it cannot ascertain which user that is. Even if the server side information is revealed, users´ privacy will not be disclosed, thus makes it suitable for personalized information retrieval in digital libraries.