NantonacCollaborativeFiltering-AModel-BasedApproach
Toshihiro Kamishima, Shotaro Akaho · 2010
A recommender system has to collect users ’ preference data. To collect such data, rating or scoring methods that use rating scales, such as good-fair-poor or a five-point-scale, have been employed. We replaced such collection methods with a ranking method, in which objects are sorted according to the degree of a user’s prefer-ence. We developed a technique to convert the rankings to scores based on order statistics theory. This technique successfully im-proved the accuracy of ranking recommended items. However, we targeted only memory-based recommendation algorithms. To test whether or not the use of ranking methods and our conversion tech-nique are effective for wide variety of recommenders, we apply our conversion technique to model-based algorithms.