LUCI: A Personalization Documentary System Based on the Analysis of the History of the Users Actions

摘要

With the development of Internet and storage devices, online document servers abound with enormous quantities of documents from various themes. The online search of pertinent documents is a fastidious task and search engine may overcome this difficulty. However, in such engines, each new document need must be formulated by a new request. Recently, new approaches were proposed to solve this problem by taking into account the user profile. However, these approaches dont consider the evolution in time of the document classes consulted by the user. In this paper, we propose a new approach for learning the user long-term profile for textual document filtering. In this framework, the documents consulted by the user are classified in a dynamic way, then we analyze the distribution in the time of the document classes. The approach aim is to determine, as well as possible, the document classes which interest the user. We also propose a system called LUCI, which allows an online documents personalization by using this approach. An empirical study confirms the relevance of our approach.