Social networks can capture a variety of relationships among the participants. Two of the most commonly studied are friendship and family, or kinship, ties. Most existing work studies these networks in isolation. Here, we study how these networks can be overlaid, and propose a feature taxonomy for link prediction in this setup. We show that when there are tightly-knit family groups, which we refer to as family circles, in a social network, we can improve the accuracy of our link prediction models. This is done by making use of the family-circle features based on the likely structural equivalence of participants in the families. We investigated the predictive power of overlaying friendship and family ties on a trio of interesting real-world social networks. Our experiments demonstrate significantly higher prediction accuracy (between 15% and 30% more accurate) as compared to using more traditional features such as descriptive node attributes and structural features. The experiments also show that a combination of all three types of attributes results in the best precision-recall trade-off.

@InProceedings{zheleva:snakdd08,
  author       = "Zheleva, Elena and Getoor, Lise and Golbeck, Jennifer and Kuter, Ugur",
  title        = "Using Friendship Ties and Family Circles for Link Prediction",
  booktitle    = "2nd ACM SIGKDD Workshop on Social Network Mining and Analysis (SNA-KDD)",
  year         = "2008",
}