First of all I greet you and to my knowledge and I hope I'm wrong for your benefit, one of the most recent references is the thesis of Christian Conte 2014. However there is a new comprehensive approach that addresses the problem and uses a nouvaux concepte'Systems one Graphs'. Here is link attached files.
-On the Identification of Polytrees of Linear Dynamic Systems with Hidden Nodes.
Keywords: Systems on Graphs, System Identification, Networked Control Systems
Abstract: This article considers the problem of reconstructing the interconnection structure of a polytree network when only non-invasive observations are available and in the potential presence of hidden nodes. Previous results were limited to tree networks with single roots. Polytrees (directed trees with multiple roots) instead represent a much richer and more expressive class of models than rooted trees. Thus the extensions obtained in this article provide a significant methodological improvement with applications to several areas such as adaptive bitrate estimation, leader detection, and disease spreading.
-Graphoid-based methodologies in modeling, analysis, identification and control of networks of dynamic systems
The conceptual and methodological gap between graphical models of random variables and networks of dynamic systems, creating a single unified language and theoretical framework for these two different classes.
Graph theory has a widespread use as a modeling tool in science. Graphs are commonly used to explain complex multivariate phenomena in terms of correlation, statistical dependence and also causal connectivity among their different parts.