A user-profile-based friendship recommendation solution in social networks

Abstract

Today, OSNs (online social networks) have become a popular communication platform, and a wide range of applications are considered for them. One fundamental relationship among members of a social network is friendship. Friend matching is a reasonable way to suggest people to each other. In this paper, we propose a model for measuring the similarities of two members based on information contained in their profiles. Unlike other profile-matching models, we do not assign equal weights to different items in users’ profiles to measure their similarity. In the first phase of the proposed scheme, we develop a mining model to discover the actual degree of influence of different factors that affect the formation of friendships. We extract this information from a real online social network. Then, based on the analysis results of this large dataset, matching and recommender systems are designed. The experimental results are encouraging and show that the proposed model significantly outperforms its counterparts.

An efficient hybrid similarity measure based on user interests for recommender systems

“…Currently, recommender systems exist in new domains. Examples include Mazhari et al (), who proposed a friendship recommender system for social networks on the basis of user profiles. The recommender system uses a similarity measure to determine the degree of similarity between two users on the basis of their profiles.…”

“…Recommender systems are used to suggest items to users based on their interests. They have been used in various domains such as market recommenders (Wei, Moreau, & Jennings, ), book recommenders (Mooney, Bennett, & Roy, ), article recommenders (Resnick, Iacovou, Suchak, Bergstrom, & Riedl, ), and social network friendship recommenders (Mazhari, Fakhrahmad, & Sadeghbeygi, ). In this paper, we focus on movie recommenders, where the items are movies and the users are online store clients.…”

User Similarity Determination in Social Networks

“…Social networks have become an integral, part of our daily lives, enabling us to keep in touch with our friends and families [1]. The existing models incorporate either node profile attributes or link formation information or both [2]. Most of them represent the members and relationships among them in the form of a mathematical graph with nodes as vertices and ties as edges.…”

“…A hybrid similarity measure that combines network similarity with node profile similarityhas been proposed in reference [5].A brief survey in reference [6] illustrates the variety of similarity measures developed for social networks and the difficulty of selecting a similarity measure for problems such as link prediction or community detection. The friendship recommendation system proposed in reference [2] employed different similarity metrics for various features. When applied to real world social network data, it suggested that interest, age, gender, and location contribute to friendship formation.…”

“…Lastly, we compared the results of both charts. A high-level diagram of the whole course of action is shown in Figure 4 [2]. The adjacency matrix for edge weight measure is tabulated below.…”