Applying SVD on Generalized Item-based Filtering

Abstract

In this paper we examine the use of a matrix factorization technique called Singular Value Decomposition (SVD) along with demographic information in Item -Based Collaborative Filtering. After a brief introduction to SVD and to some of its previous applications in Recommender Systems, we proceed with the presentation of two distinct but related algorithms. The first algorithm uses SVD in order to reduce the dimension of the active item's neighborhood. The second algorithm initially enhances Item -based Filtering with demographic information and then applies SVD at various points of the filtering procedure. The presentations of both algorithms include a detailed step-by-step description and a series of experiments. In both cases the results show that a reduction in the dimension of the item neighborhood via SVD, either by itself or combined with the usage of relevant demographic information, is promising, since it does not only tackle some of the recorded problems of Recommender Systems, but also assists in increasing the accuracy of systems employing it.