Personalized Text Categorization Using a MultiAgent Architecture

In this paper, a system able to retrieve contents deemed relevant for the users through a text categorization process, is presented. The system is built exploiting a generic multiagent architecture that supports the implementation of applications aimed at (i) retrieving heterogeneous data spread among different sources (e.g., generic html pages, news, blogs, forums, and databases); (ii) filtering and organizing them according to personal interests explicitly stated by each user; (iii) providing adaptation techniques to improve and refine throughout time the profile of each selected user. In particular, the implemented multiagent system creates personalized press-revies from online newspapers. Preliminary results are encouraging and highlight the effectiveness of the approach.