Machine Learning Models for Predicting the Quality Factor of FSO Systems with Multiple Transceivers

Free space optical (FSO) communication is a promising solution to deliver the last mile communication and to guarantee a high data rate. However, the performance of FSO links can be significantly degraded by adverse weather conditions. Recently, machine learning algorithms (MLAs) have emerged for robust prediction to optimize the network performance. In this work, the Quality factor (Q) of FSO systems is estimated by means of four MLA models, namely, multi-linear regression, support vector regression, decision tree regression, and random forest regression. The synthetic data is used for training and testing these MLAs models, and several atmosphere conditions are considered with multiple transceivers FSO link system. The results of decision tree and random forest models demonstrated high coefficient of determination (R
²
) and low mean square error (MSE) as compared to the other models.