Collective Classification in Network Data

Networks have become ubiquitous. Communication networks, financial transaction networks, networks describing physical systems, and social networks are all becoming increasingly important in our day-to-day life. Often, we are interested in models of how objects in the network influence each other (e.g., who infects whom in an epidemiological network), or we might want to predict an attribute of interest based on observed attributes of objects in the network (e.g., predicting political affiliations based on online purchases and interactions), or we might be interested in identifying important links in the network (e.g., in communication networks). In most of these scenarios, an important step in achieving our final goal, that either solves the problem completely or in part, is to classify the objects in the network. Given a network and an object o in the network, there are three distinct types of correlations that can be utilized to determine the classification or label of o: