Title :
Fuzzy-Bayesian network approach to genre-based recommender systems
Author :
Ashkezari-T, Soheila ; Akbarzadeh-T, Mohammad-R
Author_Institution :
Dept. of Comput. Eng., Ferdowsi Univ. of Mashhad, Mashhad, Iran
Abstract :
The World Wide Web has created a new media for mass marketing that can also be highly customized to online customers´ needs and expectations. Recommender Systems (RS) play an important role in this area. Here, we aim to establish a genre-based collaborative RS to automatically suggest and rank a list of appropriate items (movies) to a user based on the user profile and the past voting patterns of other users with similar tastes. The contribution of this paper is using genre based information in a hybrid fuzzy-Bayesian network collaborative RS. The interest to the different genres is computed based on a hybrid user model. The similarity of like-minded users according to the fuzzy distance and also Pearson correlation coefficient is involved in a Bayesian network.
Keywords :
Bayes methods; Internet; fuzzy set theory; recommender systems; Pearson correlation coefficient; World Wide Web; fuzzy distance; fuzzy-Bayesian network approach; genre-based recommender system; hybrid user model; mass marketing; movies ranking; movies suggestion; past voting patterns; user profile; Collaboration; Computational modeling; Equations; Mathematical model; Motion pictures; Recommender systems;
Conference_Titel :
Fuzzy Systems (FUZZ), 2010 IEEE International Conference on
Conference_Location :
Barcelona
Print_ISBN :
978-1-4244-6919-2
DOI :
10.1109/FUZZY.2010.5584250