Using Weighted Interaction Metrics for Link Prediction in a Large Online Social Network

Abstract

There has been a considerable amount of recent research on the link prediction problem, that is, the problem of accurately predicting edges that will be established between actors in a social network in a future period. With the cooperation of the provider of a German social network site (SNS), we aim to contribute to this line of research by analyzing the link formation and interaction patterns of approximately 9.38 million members of one of the largest German online social networks (OSN). It is our goal to explore the value of users’ interaction frequencies for link prediction based on metrics of local structural similarity. Analyzing a random sample of the network, we found that only a portion of the network is responsible for most of the activity observed: 42.64 % of the network’s population account for all observed interactions and 25.33 % are responsible for all private communication. We have also established that the degree of recent interaction is positively correlated with imminent link formation – users with high interaction frequencies are more likely to establish new friendships. The evaluation of our link prediction approach yields results that are consistent with comparable studies. Traditional metrics seem to outperform weighted metrics that account for interaction frequencies. We conclude that while weighted metrics tend to predict strong ties, users of SNS establish both strong and weak ties. Our findings indicate that members of an SNS prefer quantity over quality in terms of establishing new connections. In our case, this causes the simplest metrics to perform best.