Detecting Anomalies in the Optical Layer Using Unsupervised Machine Learning

We propose an unsupervised machine learning (ML) approach using field data for the detection of optical layer anomalies. We show how multivariate ML models can forecast hard failures by detecting soft failures.