A recommendation scheme by user preference to components
Zemeng Feng Zemeng Feng, Lifang Wu, Yuchen Jing Yuchen Jing, Dan Wang, H. Zhang, Cheng Zhang · 2015
Recommender system has been more and more popular no matter in e-commerce or social-multimedia. In this paper, we analyze that in most cases a person's preference to an item is influenced by his/her preference to the components of the item. In conclusion, taking food recommendation for example, we propose a recommendation scheme to item-components-item. We firstly collect user ratings to dishes, and then we decompose the dish ratings to the ingredient ratings. Finally the rating for a new dish is estimated by the user ratings to the ingredients. The experimental results indicate that the proposed approach can achieve recipe recommendation more efficiently.