Network Traffic Monitoring Using Poisson Dynamic Linear Models
- Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
In this article, we discuss an approach for network forensics using a class of nonstationary Poisson processes with embedded dynamic linear models. As a modeling strategy, the Poisson DLM (PoDLM) provides a very flexible framework for specifying structured effects that may influence the evolution of the underlying Poisson rate parameter, including diurnal and weekly usage patterns. We develop a novel particle learning algorithm for online smoothing and prediction for the PoDLM, and demonstrate the suitability of the approach to real-time deployment settings via a new application to computer network traffic monitoring.
- Research Organization:
- Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
- Sponsoring Organization:
- USDOE
- DOE Contract Number:
- W-7405-ENG-48
- OSTI ID:
- 1122210
- Report Number(s):
- LLNL-TR-483318
- Country of Publication:
- United States
- Language:
- English
Similar Records
LEISURE: Load-Balanced Network-Wide Traffic Measurement and Monitor Placement
POINT: Partially Observable Imitation Network for Traffic Signal Control
Traffic Signal Control With Adaptive Online-Learning Scheme Using Multiple-Model Neural Networks
Journal Article
·
Fri Aug 02 00:00:00 EDT 2013
· IEEE Transactions on Parallel and Distributed Systems
·
OSTI ID:1122210
+1 more
POINT: Partially Observable Imitation Network for Traffic Signal Control
Journal Article
·
Fri Oct 29 00:00:00 EDT 2021
· Sustainable Cities and Society
·
OSTI ID:1122210
+2 more
Traffic Signal Control With Adaptive Online-Learning Scheme Using Multiple-Model Neural Networks
Journal Article
·
Wed Feb 09 00:00:00 EST 2022
· IEEE Transactions on Neural Networks and Learning Systems
·
OSTI ID:1122210
+1 more