BOA based SVR analysis for Predicting the Strength of Subgrade soil for Pavement Design

Abstract

This paper introduces a hybrid of the Bayesian optimization algorithm (BOA) and support vector regression (SVR) as a new modelling tool for the California Bearing Ratio (CBR) prediction of subgrade soil for pavement design. For this purpose, an experimental database was utilized to generate the hybrid BOA-SVR model of indirect estimation of the CBR using routinely collected soil properties. The database consists of 238 experimental datasets collected from soil tests carried out in the northern region of Algeria. To develop the model, all hyperparameters were optimised using the BOA technique. It was found that the average, median, standard deviation, minimum, maximum and interquartile range of the expected values of the developed hybrid model are very close to the experimental results. Results revealed that the hybrid BOA-SVR model predict the CBR of the tested subgrade soils with a coefficient of determination of 89% and mean squared error of 5.77. Comparisons with conventional and other machine learning models showed that BOA-SVR hybrid model predictions are more accurate and robust than those of other models