Reliability Analysis of Clayey Soil Slope Stability Using GMDH and RFC Soft Computing Techniques

Soil being a heterogeneous medium, predicting the stability of soil slope is a complex engineering problem due to the involvement of multiple effective attributes of soil in geotechnical behaviour. However, as comprehension of soil variability improves, deterministic methods have been replaced by probabilistic ones. This paper examines the application of two soft computing techniques, Group Method of Data Handling (GMDH) and Random Forests Classifier (RFC), to the study of reliability analysis of clayey soil slope stability. In addition, the applicability of GMDH and RFC in predicting stability of Soil Slope based on distinct soil attributes was evaluated, and model performance was evaluated using various fitness parameters such as RMSE, LMI, Bias Factor, etc. The results indicate that the GMDH model outperformed all fitness parameters, suggesting that the GMDH approach can be used as a reliable soft computing method for addressing non-linear problems, such as the stability of soil slope.