Nowadays, digital libraries are becoming popular. These libraries provide many services for their users. The information retrieval service is an important service of the digital libraries. Personalizing this service in order to better satisfy users' requirements is an approach that attract much attention of the scientific community. Many present personalized information retrieval systems re-rank results of a search engine by taking into account the similarities between these results and the user profiles to return more relevant results. However, most of these systems only use content-based approaches for this purpose. In our work, we propose to use also citation-based methods like co-citation method and bibliographic coupling method to compute document-profile similarities. We study the performance of the co-citation method with different bibliographic databases. We also use many different combination functions to combine individual scores. The proposed approaches were validated by experiments with the test collection used in INEX 2005.
