Filtragem Colaborativa Incremental para recomendações automáticas na Web

Abstract

The use of collaborative filtering recommenders on the Web is typically done in environments where data is constantly flowing and new customers and products are emerging. In this work, it is proposed an incremental version of item-based Collaborative Filtering for implicit binary ratings. It is compared with a non-incremental one, as well as with an incremental user-based approach. It is also study the use of techniques for working with sparse matrices on these algorithms. All the versions are implemented in R and are empirically evaluated on five different datasets with various number of users and/or items. It is observed that the measure of Recall used tend to improve when we continuously add information to the recommender model and that the time spent for recommendation does not degrade. Time for updating the similarity matrix (necessary to the recommendation) is relatively low and motivates the use of the item-based incremental approach.