Predicting the drag coefficient of coastal trees using Support Vector Machines and boosting ensemble models

Abstract

Abstract The effect of green belts on wave absorption is a critical aspect of coastal protection strategies. The effectiveness of green belts in wave absorption is influenced by factors such as the type of vegetation used, the density and width of the green belt, and the topography of the coastline. The current study aims to explore the performance of various intelligent tools, including SVM (Support Vector Machine), ABR (Ada Boost Regression), ETR (Extra Trees Regression), GBR (Gradient Boosting Regression), and RF (Random Forest), to forecast drag coefficients of coastal trees (C D ). In this direction, four dimensionless parameters of relative wave height (H/d), vegetation density (D), coastal shoreline slope (S), and wave propagation velocity (u/ $$\sqrt{\xi E/\rho }$$ ξ E / ρ ) were assumed as input parameters, and C D was considered as the target. To evaluate the performance of developed soft computing models, various statistical indicators and graphical plots including Violin, Tylor, and Scatter were applied. The results revealed that the ETR method outperforms existing machine learning techniques with statistical results of R2 = 0.996, RMSE = 0.003, MAE = 0.002, and SI = 0.014. In addition, the Tylor diagram indicates that the distance index obtained using the ETR model exhibited a high alignment with actual data, especially in comparison with alternative tools.

Keywords

• Coastal trees
• Drag coefficient
• Solitary breaking wave
• Machine learning