The personalized index service system in digital library

China Digital Library is a federated digital library, which is composed of library components such as: local digital libraries, publishing companies, and so on. The objective of the federated digital library is to share resources and services of its components, therefore users can access distributed information transparently. The immense distributed electronic information available in the federated digital library raises two problems: how to find all information that the user needs accurately; and how to provide information that the user is interested. The personalized index service system presented in this paper holds the promise of resolving these problems, which is composed of an index service system and a personalized service system. The former facilitates federated resource discovery. The latter learns a model of the user´s interests and uses the model to push interesting information and sort the search results, so that users are able to easily and quickly find interesting objects. The personalized service system learns the model of the user´s initial interests after the user provides his/her interests, and the system updates the model according to user´s access log and feedback to the search results