Clustering of live network alarms using unsupervised statistical models
Résumé
An unsupervised topology and time-based clustering model is proposed to regroup alarms according to their failure events. The different modes and settings of the model are assessed using topology and alarm-related data extracted from a live network as part of a field trial.
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|