Supervised and unsupervised learning for classifying changes in Optical Time Domain Reflectometer traces

Tremblay, Christine, Tabatabaei, Mina, Boertjes, David, Pei, Yinqing et Desrosiers, Christian. 2025. « Supervised and unsupervised learning for classifying changes in Optical Time Domain Reflectometer traces ». Journal of Optical Communications and Networking.
(Sous presse)

PDF Tremblay-C-2025-31032.pdf - Version acceptée Accès restreint à : Administrateur seulement jusqu'au 25 juin 2027. Licence d'utilisation : Tous les droits réservés aux détenteurs du droit d'auteur. Télécharger (917kB) | Demande de copie

Résumé

Global telecommunications heavily rely on optical fibers as the foundation of their network infrastructure, making it imperative for network operators to ensure their dependability. Traditional Optical Time Domain Reflectometer (OTDR) focuses on event detection, but in-service measurements can detect the interactions of distributed effects such as fiber loss, Raman amplification, Stimulated Raman Scattering (SRS) and channel loading. This research paper demonstrates the effectiveness of supervised and unsupervised learning models in accurately categorizing changes observed in in-service OTDR traces. Among the supervised models tested, the Multilayer Perceptron (MLP) exhibited superior performance with a classification accuracy of 0.891 with multiple-effect data, surpassing the Random Forest (RF) and Convolutional Neural Network (CNN). Clustering models were also explored, focusing on single-effect data; the best result was obtained using the Gaussian Mixture Model (GMM), achieving a Normalized Mutual Information (NMI) of 0.663 and an Adjusted Rand Index (ARI) of 0.52.

Type de document:	Article publié dans une revue, révisé par les pairs
Professeur:	Professeur Tremblay, Christine Desrosiers, Christian
Affiliation:	Génie électrique, Génie logiciel et des technologies de l'information
Date de dépôt:	23 juin 2025 13:32
Dernière modification:	25 juin 2025 19:40
URI:	https://espace2.etsmtl.ca/id/eprint/31032

Actions (Authentification requise)

Dernière vérification avant le dépôt