Stability assessment of slopes subjected to circular-type failure using tree-based models
Vahid Amirkiyaei, Ebrahim Ghasemi
Abstract
In this paper, two tree-based models were constructed for stability assessment of slopes subjected to circular failure mode. The models include: (1) a regression tree based on M5P algorithm for prediction of slope safety factor, (2) a classification tree based on J48 algorithm for prediction of slope stability status. For constructing and validating these models, a database including 87 slope cases was used. The models evaluate the slope stability based on six major contributing parameters including slope height, slope angle, unit weight, cohesion, internal friction angle and pore pressure ratio. In validation phase of M5P model, root mean square error (RMSE), variance account for (VAF), and coefficient of determination (R2) indices were obtained as 0.09, 88.05% and 0.902, respectively. Furthermore, the accuracy of J48 model was obtained as 92.31%. The results illustrate that the developed tree models are very effective and robust tools for assessment of slope stability.