Prediction of soil unconfined compressive strength using Artificial Neural Network model

The main objective of the present study is to apply Artificial Neural Network (ANN), which is one of the most popular machine learning models, to accurately predict the soil unconfined compressive strength (q(u)) for the use in designing of foundations of civil engineering structures. For the development of model, data of 118 soil samples were collected from Long Phu 1 power plant project, Soc Trang Province, Vietnam. The database of physicomechanical properties of soils was prepared for the model study, where 70% data was used for the training and 30% for the testing of the model. Standard statistical indices, namely Root Mean Squared Error (RMSE) and Pearson Correlation Coefficient (R) were used in the validation of the model's performance. In addition, Partial Dependence Plots (PDP) was used to evaluate the importance of the input variables used for modeling. Results showed that the ANN model performed well for the prediction of the q(u) (RMSE = 0.442 and R = 0.861). The PDP analysis showed that the liquid limit is the most important input factor for modeling of the q(u). The present study demonstrated that the ANN is a promising tool that can be used for quick and accurate prediction of the q(u), which can be used in designing the civil engineering structures like bridges, buildings, and powerhouses.