ABSTRACT
In this study, the slip surface of the slope was predicted using several machine learning algorithms multilayer perceptron (MLP), XGBoost, and support vector machine (SVM). The slope used for the study is in the form of a simple finite slope, and the soil constituting the slope is sandy soil containing fine content. In performing the slope stability analysis, groundwater conditions were considered, and the influence of matric suction in unsaturated soils was also considered. The results of slope stability analysis using the limit equilibrium method were derived by applying the results of analysis on groundwater conditions using FEM. Using the Morgenstern-Price method, which can consider both force and moment equilibrium, a total of 42,000 data were produced and used for model training. In order to quantify the performance of the model in predicting the slip surface, the factor of safety of the slip surface obtained using the limit equilibrium method and the machine learning algorithm was compared.
