Structure Prediction in Uncertain Temporal Networks

Abstract

One of the tasks of network analysis is then to identify the relationships between actors in a network and their dynamic behavior using a network model with a structure that reflects best the true state of the modeled reality. Structure prediction for static networks is not a simple task. It depends on data quality. However, for networks, we are discussing in this paper, it is a computationally difficult issue. We deal with networks, which are both time-varying (temporary) and uncertain. In this paper, we use belief function theory for the modelling of uncertainty. Based on an estimate of interactions of groups of actors in a network, we estimate the structure of the modeled temporary network. Experimental results show that our method can predict the structure of indefinite time networks.