Performance estimation of next generation passive optical networks stage 2 with machine learning techniques for 5G and beyond

Study offers the idea of utilizing machine-learning (ML) to forecast performance of a 50G-WDM-PON based on dual-parallel Mach-Zehnder-Modulator. Millimeter wave-over-fiber is also introduced with dual-parallel MZM based 50G-WDM-PON network by combining the benefits of millimeter wave and fiber-optic. Machine learning uses data-driven algorithms to extract patterns and relationships from previous network performance data. The numerical simulation is investigated with machine learning model to predict the performance of the signal in terms of Q-factor and error rate. ML model provides good accuracy of greater than 75%. Only one logistic model offers less than 90%. Findings show successful performance parameters using ML.