Recommendation Process in SR1 Web Document Recommender System

Abstract

This paper presents a recommender system for web documents (given as bookmarks). The system uses for classification a combination of content, event and collaborative filters and for recommendation a modified Pearson-r algorithm. The algorithm for recommendation is using not only the correlation between users but also the similarity between classes. Some experimental results that support this approach are also presented.

Keywords

• web document recommender system
• content filtering
• collaborative filtering
• vector space model