Machine learning-based regression models for predicting signal quality of dense wavelength division multiplexing (DWDM) optical communication network

摘要

Over the years, optical communication systems have been a significant sourceof fast and secure communication. However, factors like noise and mitigationerror can degrade the bit error rate (BER) and quality factor (Q factor) of opticalcommunication systems. Predicting the optimal threshold, Q factor, andBER is usually a difficult task. Therefore, in this paper, machine learningbasedlinear regression, least absolute shrinkage and selection operator(LASSO) regression, and Ridge regression have been used for a dense wavelengthdivision multiplexing (DWDM)-based optical communication networkto predict the signal quality. These techniques have been used to predict thedesired BER, Q factor, threshold, and eye height of the system. To demonstratethis research concept, a DWDM-based optical communication network of50 km length is designed and simulated using Optisystem-14.0. After datapreparation, regression models have been developed and validated throughdiagnostic plots. Results show that mean square error (MSE) has a significantdecline with an increase in the number of epochs for all four models. LASSOand Ridge regression have effectively resolved the issue of overfitting, whichoccurred in the linear regression case. Furthermore, the mean MSE plotproved the significant reduction of mean MSE in the case of LASSO regression.Results show that min BER for LASSO regression came out to be 173,627.14,providing a robust and cost-efficient process.