Abstract
In this paper, we consider principal component analysis (PCA) in decomposable Gaussian graphical models. We exploit the prior information in these models in order to distribute PCA computation. For this purpose, we reformulate the PCA problem in the sparse inverse covariance (concentration) domain and address the global eigenvalue problem by solving a sequence of local eigenvalue problems in each of the cliques of the decomposable graph. We illustrate our methodology in the context of decentralized anomaly detection in the Abilene backbone network. Based on the topology of the network, we propose an approximate statistical graphical model and distribute the computation of PCA.
| Original language | English |
|---|---|
| Pages (from-to) | 4369-4377 |
| Number of pages | 9 |
| Journal | IEEE Transactions on Signal Processing |
| Volume | 57 |
| Issue number | 11 |
| DOIs | |
| State | Published - 2009 |
| Externally published | Yes |
Bibliographical note
Funding Information:Manuscript received August 19, 2008; revised May 04, 2009. First published June 19, 2009; current version published October 14, 2009. The associate editor coordinating the review of this manuscript and approving it for publication was Prof. C. Guillemot. The work of A. Wiesel was supported by a Marie Curie Outgoing International Fellowship within the 7th European Community Frame-work Programme. This work was also partially supported by Air Force Office of Scientific Research under Grant FA9550-06-1-0324.
Keywords
- Anomaly detection
- Graphical models
- Principal component analysis