Investigating the competency of some soft computing techniques for prediction of selected lateritic soils’ strength from index properties

This study aimed to assess the ability of some soft computing techniques including ANN, M5P and RF to accurately predict the strength of selected lateritic soils in Southwest Nigeria using index properties including specific gravity, linear shrinkage, liquid limit, plasticity index, fine sand content, and fines content. To achieve this goal, the experimental dataset obtained from the laboratory analyses of three hundred laterite samples taken from thirty different laterite deposits within Southwest Nigeria was divided into a model dataset and gaging dataset. The model dataset contains two hundred and forty data points, which were divided into 70% for training and 15% each for testing and validation, were used to train and compare the proposed models performance. The gaging dataset contains sixty data points, which were used to validate the proposed models against prominent existing models in the literature. The models performances were evaluated using various statistical estimators. Based on the statistical estimators, the proposed models outperformed the existing models in the literature and provided satisfactory performances, thus, they are validated. The obtained R2 value using the ANN model are 0.9967, 0.9963, 0.9989, and 0.9852 for the training, testing, validation, and gaging dataset, respectively; the R2 values of M5P models are 0.6676, 0.5501, 0.636 and 0.6727; while the predicted R2 values of RF model are 0.8346, 0.6380, 0.7564, and 0.7901. This implies that the ANN model is the most reliable estimation for the prediction of lateritic soils’ shear strength. Thus, ANN is strongly recommended for the prediction of lateritic soil strength.