Efficient Classification of Polarization Events Based on Field Measurements

• DOI: 10.1364/ofc.2020.th3d.7
• Publication Date: 2020-03-06T14:30:53Z
• Authors: Kyle Guan; Jesse E. Simsarian; Fabien Boitier; Dan Kilper; Jelena Pesic; Michael Sherman
• Keywords: Training set

We present rare-event classification of polarization transients based on field measurements with data augmentation combined robot-generated fiber-disturbance data. compare machine learning methods for accuracy and required number training sample traces.