Learning to Classify DWDM Optical Channels from Tiny and Imbalanced Data

Applying machine learning algorithms for assessing the transmission quality in optical networks is associated with substantial challenges. Datasets that could provide training instances tend to be small and heavily imbalanced. This requires applying imbalanced compensation techniques when using bina...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Paweł Cichosz, Stanisław Kozdrowski, Sławomir Sujecki
Formato: article
Lenguaje:EN
Publicado: MDPI AG 2021
Materias:
Q
Acceso en línea:https://doaj.org/article/7ed8b64142234126b3d69029095a2866
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:7ed8b64142234126b3d69029095a2866
record_format dspace
spelling oai:doaj.org-article:7ed8b64142234126b3d69029095a28662021-11-25T17:30:19ZLearning to Classify DWDM Optical Channels from Tiny and Imbalanced Data10.3390/e231115041099-4300https://doaj.org/article/7ed8b64142234126b3d69029095a28662021-11-01T00:00:00Zhttps://www.mdpi.com/1099-4300/23/11/1504https://doaj.org/toc/1099-4300Applying machine learning algorithms for assessing the transmission quality in optical networks is associated with substantial challenges. Datasets that could provide training instances tend to be small and heavily imbalanced. This requires applying imbalanced compensation techniques when using binary classification algorithms, but it also makes one-class classification, learning only from instances of the majority class, a noteworthy alternative. This work examines the utility of both these approaches using a real dataset from a Dense Wavelength Division Multiplexing network operator, gathered through the network control plane. The dataset is indeed of a very small size and contains very few examples of “bad” paths that do not deliver the required level of transmission quality. Two binary classification algorithms, random forest and extreme gradient boosting, are used in combination with two imbalance handling methods, instance weighting and synthetic minority class instance generation. Their predictive performance is compared with that of four one-class classification algorithms: One-class SVM, one-class naive Bayes classifier, isolation forest, and maximum entropy modeling. The one-class approach turns out to be clearly superior, particularly with respect to the level of classification precision, making it possible to obtain more practically useful models.Paweł CichoszStanisław KozdrowskiSławomir SujeckiMDPI AGarticlemachine learningoptical networksimbalanced dataone-class classificationScienceQAstrophysicsQB460-466PhysicsQC1-999ENEntropy, Vol 23, Iss 1504, p 1504 (2021)
institution DOAJ
collection DOAJ
language EN
topic machine learning
optical networks
imbalanced data
one-class classification
Science
Q
Astrophysics
QB460-466
Physics
QC1-999
spellingShingle machine learning
optical networks
imbalanced data
one-class classification
Science
Q
Astrophysics
QB460-466
Physics
QC1-999
Paweł Cichosz
Stanisław Kozdrowski
Sławomir Sujecki
Learning to Classify DWDM Optical Channels from Tiny and Imbalanced Data
description Applying machine learning algorithms for assessing the transmission quality in optical networks is associated with substantial challenges. Datasets that could provide training instances tend to be small and heavily imbalanced. This requires applying imbalanced compensation techniques when using binary classification algorithms, but it also makes one-class classification, learning only from instances of the majority class, a noteworthy alternative. This work examines the utility of both these approaches using a real dataset from a Dense Wavelength Division Multiplexing network operator, gathered through the network control plane. The dataset is indeed of a very small size and contains very few examples of “bad” paths that do not deliver the required level of transmission quality. Two binary classification algorithms, random forest and extreme gradient boosting, are used in combination with two imbalance handling methods, instance weighting and synthetic minority class instance generation. Their predictive performance is compared with that of four one-class classification algorithms: One-class SVM, one-class naive Bayes classifier, isolation forest, and maximum entropy modeling. The one-class approach turns out to be clearly superior, particularly with respect to the level of classification precision, making it possible to obtain more practically useful models.
format article
author Paweł Cichosz
Stanisław Kozdrowski
Sławomir Sujecki
author_facet Paweł Cichosz
Stanisław Kozdrowski
Sławomir Sujecki
author_sort Paweł Cichosz
title Learning to Classify DWDM Optical Channels from Tiny and Imbalanced Data
title_short Learning to Classify DWDM Optical Channels from Tiny and Imbalanced Data
title_full Learning to Classify DWDM Optical Channels from Tiny and Imbalanced Data
title_fullStr Learning to Classify DWDM Optical Channels from Tiny and Imbalanced Data
title_full_unstemmed Learning to Classify DWDM Optical Channels from Tiny and Imbalanced Data
title_sort learning to classify dwdm optical channels from tiny and imbalanced data
publisher MDPI AG
publishDate 2021
url https://doaj.org/article/7ed8b64142234126b3d69029095a2866
work_keys_str_mv AT pawełcichosz learningtoclassifydwdmopticalchannelsfromtinyandimbalanceddata
AT stanisławkozdrowski learningtoclassifydwdmopticalchannelsfromtinyandimbalanceddata
AT sławomirsujecki learningtoclassifydwdmopticalchannelsfromtinyandimbalanceddata
_version_ 1718412294650593280