As the use of online services has increased,many online vendors and service providers have attempted to improve services by personalizing the users web experience.This personalization is often accomplished by "user modeling," or an attempt to rate the relevance to a user of items in the system in order to improve search effectiveness.A review of recent literature shows that user modeling benefits both the service provider and the user by facilitating interactions between the two.Relevance ratings for particular users are very useful in personalization of search results.However,determining relevance for a user is generally accomplished by the collection of the users personal information,and the collecting and processing of this information can conflict with user privacy.Unless a vendor can guarantee the protection of personal information,users will probably withhold some information and personalization will not be as effective as it could be.