Dynamic drift-adaptive ensemble-based quality of transmission classification framework in OTN - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Dynamic drift-adaptive ensemble-based quality of transmission classification framework in OTN

Résumé

We introduce Dynamic Drift-Adaptive Ensemble-based Quality of Transmission (QoT) Classification Framework (DAEQoT) to effectively verify the feasibility of a lightpath while maintaining high prediction accuracy in a dynamic Optical Transport Network (OTN) scenario. This framework adopts the Early Drift Detection Method (EDDM) to identify any significant increase in the prediction error. Moreover, Ensemble Learning is implemented to enhance the accuracy of the Master QoT model by combining other Supportive QoT classifiers when an early drift warning is reported. Thus, prediction error and complete model retraining are mitigated. DAEQoT achieves the best accuracy of 98.56% and reduces the execution time up to 52.81% compared to state-of-the-art offline and online machine learning approaches.
Fichier principal
Vignette du fichier
publi-7381.pdf (1.16 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04392273 , version 1 (13-01-2024)

Identifiants

Citer

Huy Quang Tran, Javier Errea, Van-Quan Pham, Dominique Verchere, Adlen Ksentini, et al.. Dynamic drift-adaptive ensemble-based quality of transmission classification framework in OTN. 2023 23rd International Conference on Transparent Optical Networks (ICTON), Jul 2023, Bucharest, France. pp.1-4, ⟨10.1109/ICTON59386.2023.10207229⟩. ⟨hal-04392273⟩
4 Consultations
12 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More