Prediction of unconfined compressive strength of cement–fly ash stabilized soil using support vector machines

In the current study, an effort is made to stabilize intermediate plastic clay using various dosages of cement and fly ash, and to develop an SVM-based machine learning model to predict the unconfined compressive strength of the stabilized soil. The materials used in the study were subjected to basic engineering tests such as microstructural characterization and unconfined compressive strength (UCS) test. The UCS results of the samples show a steady increase in strength value with the increase in the curing time and the cement content. Support vector machine (SVM)-based UCS prediction models with different kernel functions: linear, radial bias function, and POWER, were also developed and were compared with a multiple regression model. The training dataset of the study contained 72 data points and the testing dataset contained 18 data points. A tenfold cross validation was also employed to validate the training results. The study showed that the SVM model developed with the RBF kernel function outperformed all the other models with a R 2 value of 0.94 in training and 0.958 in testing. The sensitivity analysis of the input parameters showed that the cement(%) has the maximum effect (0.75) on the prediction of UCS by the RBF kernel-based prediction model. © 2023, The Author(s), under exclusive licence to Springer Nature Switzerland AG.