The DELOS Association for Digital Libraries has been established in order to keep the "DELOS spirit" alive by promoting research activities in the field of digital libraries. More info...
2007-06-08: Second Workshop on Foundations of Digital Libraries
The 2nd International Workshop on Foundations of Digital Libraries will be held in Budapest (Hungary) on 20 Septemeber 2007, in conjunction with the 11th European Conference on Research and Advanced Technologies for Digital Libraries (ECDL 2007). Event website
DL Events
January 24-25, 2008 - Padova, Italy
4th Italian Research Conference on Digital Library Systems Event website
December 5-7, 2007 - Pisa, Italy
Second DELOS Conference on Digital Libraries Event website
Contact Point: Giovanni Semeraro (
) University of Bari, Italy
Technical Contact Points: Pierpaolo Basile (
) and Marco Degemmis (
)
Type of software: PROTOTYPE
Descriptive Keywords: Content-based Recommender System, User Profiling System
Potential Use and Applications: The main target of personalization in a virtual information and knowledge environment is the reduction of information overload. User modeling and tracking activities provide the basis for a wide range of services that reuse the semantic information describing properties of the user. Such personalization services contribute to a more targeted information access. Typical applications of user profiles in an information environment are information filtering, personalized information recommendations, and targeted notification about changes in the information space. ITR system could be used for developing intelligent recommendation services in which items to be recommended are described by using text.
General Description: ITem Recommender (ITR) is a content-based profiling system able to infer user profiles from textual documents rated by users according to their interests. User profiles are exploited in a recommendation process to filter and to suggest new documents to the users.
The system implements a naïve Bayes text categorization algorithm and it is able to classify documents as interesting or uninteresting for a specific user by exploiting a probabilistic model learned from training examples, the documents rated by that user in the past. ITR represents documents by using senses corresponding to concepts identified from words in the original text through an automated Word Sense Disambiguation procedure. The final outcome of the learning process is a probabilistic model of the user interests. The model is used as a personal profile including those concepts that turn out to be most indicative of the user’s preferences, according to the value of the parameters of the model.