A Random Forest Algorithm Combined with Bayesian Optimization for Atmospheric Duct Estimation

• Published in: Remote sensing (Basel, Switzerland) Vol. 15; no. 17; p. 4296
• Main Authors: Yang, Chao, Wang, Yulu, Zhang, Aoxiang, Fan, Hualei, Guo, Lixin
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.09.2023
• Subjects: Accuracy; Algorithms; Approximation; Artificial intelligence; Atmospheric boundary layer; atmospheric duct estimation; Bayesian analysis; Bayesian theory; Communication; Communications systems; Comparative analysis; Decision trees; Environmental aspects; Genetic algorithms; Identification and classification; Lower atmosphere; Machine learning; Mathematical models; Mathematical optimization; Measurement; Neural networks; Optimization; orthogonal experiment design; parabolic equation; Parameters; Performance evaluation; prediction; radar; radar sea clutter; Radio frequency; random forest; Remote sensing; Sea level; Telecommunication; China
• ISSN: 2072-4292; 2072-4292
• DOI: 10.3390/rs15174296

Inversion of atmospheric ducts is of great importance in the field of performance evaluation for radar and communication systems. Since the model parameters in machine learning play a crucial role in prediction performance, this paper develops a random forest (RF) model integrated with Bayesian optimization (BO) called BO-RF for atmospheric duct prediction, and the BO is adopted to determine appropriate model parameters during the training process. In addition, the K-fold cross-validation (CV) method is also incorporated into the model to obtain the best model partition and overcome the overfitting problem. To test the performance of the proposed model, the results obtained by the BO-RF are compared with other commonly used methods, such as classical RF, extreme gradient boosting (XGBoost) with/without BO, and K-nearest neighbor (KNN) with/without BO. Comparisons demonstrate that BO-RF has the best accuracy and anti-noise ability for the estimation of duct parameters.