Applying Item-based and User-based collaborative filtering on the Netflix data

The GroupLens research lab introduced Item-Based collaborative filtering [3] as an approach for providing personalised recommendations superior to User-based collaborative filtering. User-based collaborative filtering [5] gives personalised recommendations by finding similiar users. Item-Based collaborative filtering recommends similiar items. Both approaches have been compared [3]. However, the applied testing procedure did not employ equal conditions for both approaches. The aim of this report is to give an evaluation on both approaches by employing a fair testing procedure on data provided by Netflix [2]. Test results and their dependency to the employed algorithms are interpreted.