Improving individual predictions using social networks assortativity

Social networks are known to be assortative with respect to many attributes such as age, weight, wealth, ethnicity and gender. Independently of its origin, this assortativity gives us information about each node given its neighbors. It can thus be used to improve individual predictions in many situations, when data are missing or inaccurate. This work presents a general framework based on probabilistic graphical models to exploit social network structures for improving individual predictions of node attributes. We quantify the assortativity range leading to an accuracy gain. We also show how specific characteristics of the network can improve performances further. For instance, the gender assortativity in mobile phone data changes significantly according to some communication attributes