Friend recommendation with a target user in social networking services

• Published in: 2015 31st IEEE International Conference on Data Engineering Workshops pp. 235 - 239
• Main Author: Sundong Kim
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.04.2015
• Subjects: Algorithm design and analysis; Approximation algorithms; Approximation methods; Feeds; Greedy algorithms; Monte Carlo methods; Social network services
• DOI: 10.1109/ICDEW.2015.7129582

Friend recommendation is one of the primary functions in social networking services. Suggesting friends has been done by calculating node-to-node similarity based on topological location in a network or contents on a user's profile. However, this recommendation does not reflect the interest of the user. In this paper, we propose a friend recommendation problem in which the source user wants to get more attention from a special target. The goal of our friend recommendation is finding a set of nodes, which maximizes user's influence on the target. To deliver this problem, we introduce information propagation model on online social networks and define two measures: influence and reluctance. Based on the model, we suggest an IKA(Incremental Katz Approximation) algorithm to effectively recommend relevant users. Our method is compared with topology-based friend recommendation method on synthetic graph datasets, and we show interesting friend recommendation behaviors depending on the topological location of users.