User Recommendation in Microblogs based on Interaction Networks and Contents
Hiroki Okamoto, Masashi Toyoda, Kitsuregawa Masaru · IEICE Technical Report; IEICE Tech. Rep. · 2013
User recommendation on social networks is one of the most important service on social media. It generates more links between users, and users can communicate with more friends and efficiently correct information. User recommendation is one of the link prediction problems. There have been several methods based on graph structure and contents of users’ posts, but these features have not been well combined in previous methods. In this paper we propose two user recommendation methods that utilizes a random walk technique combined with contents similarities. We apply these methods to a Twitter dataset and compare their accuracy.