A Comparative Study on Semi-Supervised Learning Algorithms for Fault Management in Optical Networks

Authors

DOI:

https://doi.org/10.14209/jcis.2026.1

Keywords:

Fault management, Optical networks, Semi-supervised Learning, Cluster-based algorithms, Dimensionality reduction techniques

Abstract

In recent years, the demand for reliable optical networks has intensified due to the rise of bandwidth-intensive applications. Faults in these networks can degrade transmission quality, leading to packet loss and service disruptions, making effective failure management essential for maintaining network availability and compliance with Service Level Agreements (SLAs). While most contemporary approaches rely on supervised learning (SL), which requires large amounts of often scarce fault data for training, this study explores semi-supervised learning (Semi-SL) algorithms for fault management in optical networks, which rely only on normal operating condition data, thus reducing the dependency on scarce fault data. We evaluate several cluster-based algorithms and dimensionality reduction techniques using a dataset from an optical testbed. Results based on Type I/II errors show that all techniques performed with an average accuracy exceeding 90%. The Autoencoder yielded the best fault management performance (96.46% of average accuracy), followed by Gaussian Mixture Model (92.5%), Mahalanobis Squared-Distance (92.48%), Density-based spatial clustering of applications with noise (92.42%), Fuzzy C-means (92.12%), K-means (92.09%), and Principal Component Analysis (90.22%).

Downloads

Download data is not yet available.

Author Biographies

Ingrid Ramos, Federal University of Pará

She is an undergraduate student in Computer Engineering at the Federal University of Pará. She is an IT-A scholarship holder at the Laboratory of Applied Electromagnetism (LEA), where she conducts research on Machine Learning applied to anomaly detection.

Andrei Ribeiro, Federal University of Pará

PhD student in Electrical Engineering at the Graduate Program in Electrical Engineering (PPGEE) of the Federal University of Pará (UFPA). Master's degree in Electrical Engineering from PPGEE and Telecommunications Engineer from UFPA. He is a researcher at the Laboratory of Applied Electromagnetism (LEA) at UFPA, where he conducts research involving Deep Learning techniques for optical network management. He has over 3 years of experience in research projects applying Machine Learning to areas such as agriculture, healthcare, and optical networks. He worked for a short period at the Samsung RD Institute as a research and development apprentice.

Fabrício Lobato, Federal University of Pará

Bachelor's degree in Computer Science from the University of the Amazon (2005), Full License in Mathematics from the State University of Pará (2004), Master's degree in Computer Science from the Federal University of Pará (2009), and PhD in Electrical Engineering from the Federal University of Pará (2019). He was a tenured professor at the Federal University of Western Pará until 2021 and is currently an Associate Professor I at the Faculty of Computer Engineering and Telecommunications of the Institute of Technology at the Federal University of Pará (UFPA). 

Moisés F. Silva, Los Alamos National Laboratory

Received the Ph.D. degree in electrical engineering from the Federal
University of Pará, Brazil, in 2020. Currently a postdoctoral research associate at the Los
Alamos National Laboratory (LANL), as part of the Mechanical and Thermal Engineering (E1) Group, working on the development of algorithms for inverse neural rendering and
neuromorphic imaging for photo-realistic 3D modeling of scenes and monitoring of highspeed physical phenomena. Before joining LANL he has been a postdoctoral fellow at the
Scuola Superiore Sant’Anna, Italy, for one and a half years performing research activities
related to intelligent software-defined optical networks, internet of things, and network
automation. Moises has collaborated in more than 30 papers published in peer-reviewed
international journals and over 30 publications in national and international conference
proceedings, currently serving as reviewer for a few international journals. At LANL, he has
served as co-mentor for students enrolled in the Los Alamos Dynamics Summer School for a
few years, leading to papers published in conference proceedings. Also, while at LANL, he
has been granted a conference award for his work involving an LDRD project to develop an
imager-based motion identification technique from dynamic point clouds using blind source
separation and motion magnification techniques.

João CWA Costa, Federal University of Pará

He holds a degree in Electronic Engineering from the Federal University of Pará (1981), a Master's degree in Electrical Engineering - Telecommunications from the Pontifical Catholic University of Rio de Janeiro (1989), and a PhD in Electrical Engineering - Telecommunications from the State University of Campinas (1994). He worked as a telecommunications engineer at DENTEL (ANATEL) and FUNTELPA between 1981 and 1986, developing installation feasibility projects and participating in the implementation of various sound and image broadcasting systems, including FM-Cultura, TV-Cultura, and the proposal for the new Basic TV Channel Plan for the State of Pará. Since 1994, he has been a faculty member at UFPA and became a full professor in 2018, working at the Faculty of Computer Engineering and Telecommunications at the Institute of Technology. He is a permanent researcher at the Graduate Program in Electrical Engineering and contributed to the creation of telecommunications and applied computing areas. He was the coordinator of the PPGEE (Oct 2004 - Dec 2005), Director of Research, and coordinator of the PIBIC at UFPA. He led the creation of the Graduate Program in Computer Science and was a permanent professor until 2023. He served as Deputy Secretary of Development, Science, and Technology - SEDECT (Jan 2007 - Dec 2010) and participated in structuring the CT system in the State of Pará. He was the President of the Brazilian Society of Microwaves and Optoelectronics (2012-2014), Vice-President (2010-2012), and is a member of the Brazilian Telecommunications Society and IEEE. A CNPq researcher since 1994, he was the first pro tempore Vice-Rector of the Federal University of South and Southeast Pará from August 2013 to February 2016. His research focuses on communication networks, applied electromagnetism, and applied computing. He has supervised more than a hundred graduate and postgraduate students, and he is a co-author of five international patent applications. He is a founding member and the first president of iSACI (Institute for Sustainability of the Amazon with Science and Innovation) and, since January 2025, has been the president of the Guamá Foundation for Science, Technology, Innovation, and Sustainable Development.

Downloads

Published

2026-01-09

How to Cite

Ferreira de Magalhães Ramos, I. L., Nogueira Ribeiro, A., Rossy de Lima Lobato, F., Silva, M. F., & Crisóstomo Weyl Albuquerque Costa, J. (2026). A Comparative Study on Semi-Supervised Learning Algorithms for Fault Management in Optical Networks. Journal of Communication and Information Systems, 41(1), 1–15. https://doi.org/10.14209/jcis.2026.1

Issue

Vol. 41 No. 1 (2026)

Section

Regular Papers

License

Copyright (c) 2026 Ingrid Ramos, Andrei Ribeiro, Fabrício Lobato, Moisés F. Silva, João CWA Costa (Author)

Creative Commons License

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

Authors who publish in this journal agree to the following terms:

  1. Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a CC BY-NC 4.0 (Attribution-NonCommercial 4.0 International) that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this journal.
  2. Authors can enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
  3. Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) before and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).

___________

Received 2025-08-20
Accepted 2025-12-30
Published 2026-01-09