An Approach to Predict User's Interests in Web-based Educational Systems using a Collaborative Filtering Weighted Method

This paper presents an approach to predict userpsilas interests in Web-based educational systems, using the collaborative filtering methods weighted by implicit information about userpsilas navigation. To test and evaluate the proposed approach, we used a particular case study. The values were obtained by a domain specialist and we compared the approach results using a cross-validation method. The obtained results show that only the traditional prediction in collaborative filtering does not assert that the userpsilas classification is the real userpsilas relationship with the system.