| ]: Probability And Statistics Contin.qency table anal- (2007) | |||||||||||||||||
Abstract | |||||||||||||||||
| ysis Ve have developed a method for recommending items that combines content and collaborative data under a single probabifistic framework. We benchmark our algorithm against a naYve Bayes classifier on the cold-start problem, where we wish to recommend items that no one in the commu-nity has yet rated. Ve systematically explore three testing methodologies using a publicly available data set, and explain how these methods apply to specific real-world appli-cations. Ve advocate heuristic recommeuders when bench-marking to give competent baseline performance. Ve introduce a nev perfbrmance metric, the CROC curve, and demonstrate empirically that the various components of our testing strategy combine to obtain deeper understanding of the performance characteristics of recommender systems. Though the emphasis of onr testing is on cold-start recommending, our methods fbr recommending and evaluation are general. | |||||||||||||||||
Publication details | |||||||||||||||||
| |||||||||||||||||