Features
- Cover Type: Hard Cover with 225 pages
- Published by: Birkhäuser Boston
- Edition: 1st Edition December 1, 2006
- Written in: English
- ISBN 10 Number: 0817644857
- ISBN 13 Number: 978-0817644857
-
Book Dimensions:
9.3 x 6.3 x 0.6 inches
- Weighs: 1 pounds
Reader Reviews
This book principally deals with the application of graph-theoretic algorithms (matching, similarity measures, graph change detection, clustering,...) to real-world networks. The focus of the book is on the graph-theoretic tools, i.e. adapting theoretical tools to get insight into the dynamics of networks. In the first part, the authors review the state-of-the-art in network management and monitoring, and review the current work on anomaly detection. Some background on graph theory is also presented in the second chapter. In the second and main part, the authors apply in each of the 6 chapters graph-theoretic tools and concepts on a particular problem. Emphasis is put on the computational cost of the techniques. Many interesting problems for detecting abnormal changes in graph dynamics are proposed based on edit distance and graph clustering. In the third part, topological metrics are used to better understand the structure of enterprise networks. Tournament scoring is also proposed to rank the nodes by their importance in communications. The fourth and last part deals with the problem of recovering missing information from graphs. Trying to recover missing information helps to reduce the risk of false alarms in abnormal change.
Comment | |
(Report this)
Search for books at
