Support Vector Machine to Predict the Pile Settlement using Novel Optimization Algorithm

Project Immunization, like piled construction, requires considerations that make them safe during the period of operation. Pile Settlement (PS), a vital issue in projects, has attracted many regards to avoid failure before commencing employing constructions. Several factors in appraising the pile movement can assist in understanding the future of the project in the loading stage. Many intelligent strategies to mathematically compute the pile motion are employed to simulate the PS. The present study aims to use Support vector regression (SVR) to predict the settlement of piles. In addition, to improve the accuracy of the related model, two meta-heuristic algorithms have been used, including the Arithmetic Optimization Algorithm (AOA) and Grasshopper Optimization Algorithm (GOA), a hybrid format in the framework of SVR-AOA and SVR-GOA. Kuala Lumpur transportation network was chosen to investigate the pile motion according to the ground properties’ condition with SVR-AOA and SVR-GOA developed frameworks. For the evaluation of each model’s performance, five indices were employed. That, the values of RMSEs for SVR-AOA and SVR-GOA were obtained at 0.550 and 0.592, respectively, and MAE exhibited the values of 0.525 and 0.561 alternatively. The R-value for the SVR-AOA showed a desirable magnitude of 0.994, which is 0.10% higher than the SVR-GOA. Also, OBJ, including R, RMSE, and MAE , for SVR-GOA and SVR-AOA were computed at 0.541 and 0.586 mm, respectively. Models’ results have had a similar performance to estimating the PS rate.