Preference relation based collaborative filtering with graph aggregation for group recommender system
- PDF / 1,094,656 Bytes
- 15 Pages / 595.224 x 790.955 pts Page_size
- 74 Downloads / 234 Views
Preference relation based collaborative filtering with graph aggregation for group recommender system Abinash Pujahari1
· Dilip Singh Sisodia1
© Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Most of the group recommender systems (GRS) apply some aggregation strategy to the ratings given by the group members for generating recommendations. But this can be highly influenced by a few members of the group, which can lead to poor group recommendation. Further, rating based aggregation strategies do not provide efficient ranking of items. Keeping these things in mind, this paper proposes a preference relation (PR) based GRS, that uses matrix factorization (MF) for predicting unknown PRs for group members. The aggregation of preferences is done using a novel virtual user based weight aggregation strategy. The weight aggregation concept is derived from the graph aggregation process. The advantage of this process is that it does not ignore weak preferences and also contributes towards group recommendation. The proposed model is evaluated and compared using standard ranking measures for MovieLens and NetFlix datasets. Experimental results obtained using Top-K recommendation task indicates the superiority of the proposed GRS method over the others. The proposed GRS model provides the best performance when we balance the number of member in a group and the number of recommended items. Keywords Group recommendation · Preference relation · Matrix factorization · Graph aggregation
1 Introduction Recommender System(RS) [5, 40] is a useful tool which filters a large amount of information to help us find the data of our need. The information can be a set of movies, books, songs, etc. Hence, most of the websites use some kind of RS to facilitate their end-users. The use of RSs are also extended to domains like social recommendation [46], music [26] and many more. RS is one of the popular research area these days, thanks to its popularity. Generally, We have two kinds of RSs: ”Personal Recommender System” (PRS) which generates personal recommendations and ”Group Recommender System” (GRS) that provides group recommendations. Between these two, PRS is the most explored RS category. However, GRS is slowly gaining popularity in the
Abinash Pujahari
[email protected] Dilip Singh Sisodia [email protected] 1
Department of Computer Science, Engineering, National Institute of Technology Raipur, GE Road, Raipur, Chhattisgarh 492010, India
recent past [20]. Because, there are many scenarios where we need recommendations for a group rather than individuals, such as tourism and entertainment purposes. Here, we need to consider the preference of all the group members. There are two ways a RS can generate recommendations: Collaborative Filtering(CF) [27] and Content-Based Filtering(CB) [29]. CF based RS analyses user-item interaction to filter users with similar kind of preferences to a target user (for whom we want to generate recommendations). Then the RS lists out a set of items that a
Data Loading...
