Pre-, In-, and Post-Processing Class Imbalance Mitigation Techniques for Failure Detection in Optical Networks
Machine Learning
2025-07-30 v1 Signal Processing
Optics
Abstract
We compare pre-, in-, and post-processing techniques for class imbalance mitigation in optical network failure detection. Threshold Adjustment achieves the highest F1 gain (15.3%), while Random Under-sampling (RUS) offers the fastest inference, highlighting a key performance-complexity trade-off.
Cite
@article{arxiv.2507.21119,
title = {Pre-, In-, and Post-Processing Class Imbalance Mitigation Techniques for Failure Detection in Optical Networks},
author = {Yousuf Moiz Ali and Jaroslaw E. Prilepsky and Nicola Sambo and João Pedro and Mohammad M. Hosseini and Antonio Napoli and Sergei K. Turitsyn and Pedro Freire},
journal= {arXiv preprint arXiv:2507.21119},
year = {2025}
}
Comments
3 pages + 1 page for acknowledgement and references