An interactive system developed for adaptive content-based retrieval of documents belonging to Web digital libraries. The distinctive characteristic is its mechanism for the creation and management of a stereotype knowledge base, and its use for user modeling. The system helps the domain expert build the stereotypes through a k-means clustering technique which is applied to the whole document collection in an off-line phase.The user’s profile evolves over time in accordance to the user’s information needs, formulated through queries and relevance feedback allowing the user to provide an assessment of the documents retrieved by the system.The filtering system extends the traditional one based on the Vector Space model because it also takes into account the co-occurrences of terms in the computation of document relevance and involves user profiles to perform query expansion.The results of the experiments are promising, both in terms of performance and in the ability to adapt to the user’ shifting interests.
GENTILI G., MICARELLI A., SCIARRONE F.. Infoweb: An Adaptive Information Filtering System for the Cultural Heritage Domain. Applied Artificial Intelligence 17(8-9): 715-744 (2003).