Machine learning approaches to estimation of the compressibility of soft soils

The modulus of compression and coefficient compressibility soft soils are key parameters for assessing deformation geotechnical infrastructure. However, the consolidation tests used to determine these two indices time-consuming results easily heavily influenced by workmanship, testing apparatus, other factors. Therefore, it is great interest develop a simple approach accurately estimate indices. This article presents development three machine learning (ML) models—at artificial neural network (ANN), random forest model, support vector model—for mapping soils. A database containing 743 sets measured physical was adopted train validate models. To quantify model uncertainty, accuracies ML models were statistically evaluated using bias factor defined as ratio predicted showed that all accurate on average, with low dispersion in prediction accuracy. ANN found be best provides analytical form has no hidden dependency between Finally, probability distribution functions factors also determined fit-to-tail technique. this study will helpful saving cost time investigation