Computer Science · Artificial Intelligence
Attributed Graph Models: Towards the Sharing of Relational Network Data
Joseph J. Pfeiffer, Sebastián Moreno, Timothy La Fond, Jennifer Neville, Brian Gallagher
See on the mapAbstract
The growth of the internet has created large scale col-lections of relational data. In these cases, datasets contain relationships between the items or individuals that are be-ing modeled – e.g, hyperlinks connect webpages on the in-ternet, while friendships (Facebook), followers (Twitter) or messages (Email) form between individuals in social net-works. Individuals connected through these relationships exhibit relational correlation, or a statistical dependence of their attributes [2]. Modeling these relationships can give better predictions about users, or a better understanding of the underlying social processes. The field of Statistical Relational Learning (SRL) utilizes these relational connections to collectively predict the un-known labels in the network, with resulting methods able to largely outperform traditional independent learning meth-