Social Network Influence Ranking via Embedding Network Interactions for User Recommendation.

Within social networks user influence may be modelled based on user interactions. Further, it is typical to recommend users to others. What is the role of user influence in user recommendation? In this paper, we first propose to use a node embedding approach to integrate many types of interaction into embedded spaces where we then define a novel closeness measure to quantify the closeness of users based on interactions. We then propose a new influence ranking algorithm based on PageRank by incorporating the closeness measure into the ranking mechanism. We evaluate our algorithm, EIRank, using a dataset collected from Twitter. Our experimental results show that our algorithm measures user influence better by way of a user recommendation task, where our algorithm outperforms TwitterRank across a range of experimental network settings.