Assistant Professor Dr. Derek Doran recently published a book on the topic of network role mining. This is the task of classifying nodes in a network with labels representing their contribution or function to the structure of and operation of a complex system. For example, in an online social network describing relationships in an online forum, there may be “lurkers” who never post, “trolls” who try to instigate arguments, and “contributors” who post often, reply to others, and make the forum a useful source of information. The methods often assume that the role classes are hidden, and then find a most probable assignment of roles with some sociological, probabilistic, or machine learning method.
Check it out at Springer.com! More information about the latest research going on in his lab can be found here!
