A new approach to multimedia information filtering based on its structure

Abstract

In information filtering systems, the documents are sequentially presented to the users based on the user relevance values. This paper argues that the presented documents should be both important and relevant to the users. Based on this view, a structure based filtering framework is described, which considers the importance and relevance values of documents. An approach, which is used to calculate the importance values of documents, is proposed to improve the representation of user profiles. The experiment results and discussion are given.