|
Abstract:
|
The statistical monitoring of networks is a new and fruitful area of research. In this talk, we investigate a relatively new mapping scheme between graphs and time series and explore its application to the general network monitoring and surveillance problem. In particular, we investigate how shifts in specific graph features, e.g., community structure, graph hierarchy, etc., are reflected in the mapping from network to time series, and then use this information to develop a tractable statistical process monitoring strategy useful in monitoring for changes in these graph features.
|
