A Machine-Learning Based Predictive Model for Liquefaction Ejecta Manifestation in Stratigraphically Variable Sites

This project presents a machine learning based probabilistic predictive model for estimating the potential severity of free-field surface ejecta in stratigraphically variable and interbedded deposits as a measure of liquefaction manifestation and damage. Results from the numerical database used to create the model are also provided which includes 19,760 fully-coupled, nonlinear, dynamic, effective stress, finite element analyses of free-field soil columns in 3D subject to 1D vertical propagation of horizontal shear waves.