|
Activity Number:
|
324
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statisticians in Defense and National Security
|
| Abstract - #306651 |
|
Title:
|
Network Traffic Profiling
|
|
Author(s):
|
Patricia H. Carter*+
|
|
Companies:
|
Naval Surface Warfare Center
|
|
Address:
|
Code B21, Building 1470, Room 2311, Dahlgren, VA, 22448,
|
|
Keywords:
|
network traffic ; document space ; clustering ; graph models ; anomaly detection ; classification
|
|
Abstract:
|
The goal of this work is to characterize normal versus non-normal computer network traffic state at different temporal and spatial granularities. A graph theoretic representation is used in which the nodes are an abstraction of all traffic over a unit of time; this abstraction is the frequencies of groups of network connections of various types. The weight of an edge between two nodes is a measure of similarity based on document similarity measures. We apply techniques from document clustering and classification and scan statistics in this setting.
|
