Bayesian Dynamic Modeling and Analysis of Streaming Network Data

Authors: 
Xi Chen, Kaoru Irie, David Banks, Robert Haslinger, Jewell Thomas & Mike West
Duke University, MaxPoint Interactive Inc

Jun 11 2016

Original version: July 2015

Traffic flow count data in networks arise in many applications, such as automobile or aviation transportation, certain directed social network contexts, and internet studies. Using an example of internet browser traffic flow through domains of an international news website, we present Bayesian analyses of two linked classes of models which, in tandem, allow fast, scalable and interpretable Bayesian inference. We first develop flexible state-space models for streaming count data, able to adaptively characterize and quantify network dynamics effectively and efficiently in real-time. We then use these efficiently implemented models as emulators of more structured, time-varying gravity models that allow closer and formal dissection of network dynamics. This yields interpretable inferences on traffic flow characteristics, and on dynamics in interactions among network nodes. Bayesian model monitoring theory defines a strategy for sequential model assessment and adaptation in cases of signaled departures of network flow data from model-based predictions. Exploratory and sequential monitoring analyses of evolving traffic on a defined network of web domains in e-commerce demonstrate the utility of this coupled Bayesian emulation approach to analysis of streaming network count data.

Keywords: 

Bayesian model emulation, Decouple/Recouple, Dynamic network flow model, Dynamic gravity model, Monitoring and anomaly detection

Manuscript: 

PDF icon -02.pdf

BibTeX Citation: 

@ARTICLE{Chen_dynets2016,
  author = {X. Chen and K. Irie and D. Banks and R. Haslinger and J. Thomas and M. West},
  title = {Scalable {B}ayesian modeling, monitoring and analysis of dynamic network flow data},
  journal = {Technical Report, Duke University},
  year = {2016},
  note = {Original version: July 2015. Current version: Invited revision under
	review at: {\em Journal of the American Statistical Association}},
  url = {https://stat.duke.edu/research/papers/2015-02}
}