Discovery of anomalous behaviour in temporal networks

In this work, we consider the problem of detecting anomalous behaviour, and we present a novel approach allows classifying ”behaviour” either to be normal or abnormal by checking the p-value associated with the occurrence of the behaviour which is modelled following a binomial distribution within a discrete time model. The basic model for social networks is a graph. Over time, the graph underlying the communications of the social network will change, as the active members of the social network communicate. We investigate the problem of detecting anomalous behaviour by looking at how the communication evolves over time. Some prior negative qualitative information on some nodes of the network is assumed, and we propose an analysis which allows to infer a subset of nodes of the social network which might share the same negative connotation. In other words, assuming one or more members belong to some criminal organisation, we wish to investigate how many other persons belong to the same organisation.