A multiobjective genetic algorithm based hybrid recommendation approach

Personalized recommendation approaches have received much attention over the years. In this paper, we propose a hybrid recommendation approach that integrates an item-based collaborative filtering, a user-based collaborative filtering and a matrix factorization method. The approach considers the two...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:SSCI : 2017 IEEE Symposium Series on Computational Intelligence : November 27, 2017-December 1, 2017 s. 1 - 6
Hlavní autoři: Wang, Pan, Zuo, Xingquan, Guo, Congcong, Li, Ruihong, Zhao, Xinchao, Luo, Chaomin
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 01.11.2017
Témata:
hybrid algorithm
Linear programming
Matrix decomposition
Measurement
multiobjective optimization algorithms
recommendation algorithms
Sociology
Sparse matrices
Statistics
Telecommunications
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Personalized recommendation approaches have received much attention over the years. In this paper, we propose a hybrid recommendation approach that integrates an item-based collaborative filtering, a user-based collaborative filtering and a matrix factorization method. The approach considers the two objectives of recommendation's accuracy and diversity simultaneously. First, a set of items is created separately by each of the three methods. Then, items produced by the three methods are combined into a set of candidate items. Finally, a multiobjective genetic algorithm is adopted to choose a set of Pareto recommendation lists from the set. Experimental results show that the proposed approach is very effective and is able to produce better Pareto solutions than those comparative approaches.
DOI:10.1109/SSCI.2017.8285336