Machine-Learning-Based Classification Approaches toward Recognizing Slope Stability Failure

Moayedi Moayedi,Kalantar Kalantar,Bui Bui,Foong Foong

doi:10.3390/app9214638

Abstract

In this paper, the authors investigated the applicability of combining machine-learning-based models toward slope stability assessment. To do this, several well-known machine-learning-based methods, namely multiple linear regression (MLR), multi-layer perceptron (MLP), radial basis function regression (RBFR), improved support vector machine using sequential minimal optimization algorithm (SMO-SVM), lazy k-nearest neighbor (IBK), random forest (RF), and random tree (RT), were selected to evaluate the stability of a slope through estimating the factor of safety (FOS). In the following, a comparative classification was carried out based on the five stability categories. Based on the respective values of total scores (the summation of scores obtained for the training and testing stages) of 15, 35, 48, 15, 50, 60, and 57, acquired for MLR, MLP, RBFR, SMO-SVM, IBK, RF, and RT, respectively, it was concluded that RF outperformed other intelligent models. The results of statistical indexes also prove the excellent prediction from the optimized structure of the ANN and RF techniques.

Highlights

The stability of local slopes is a critical issue that needs to be meticulously investigated due to their great impact on the adjacent engineering projects
In this paper, we investigated the efficiency of seven machine-learning models, namely multiple linear regression (MLR), multi-layer perceptron (MLP), radial basis function regression (RBFR), improved support vector machine using sequential minimal optimization algorithm (SMO-SVM), lazy k-nearest neighbor (IBK), random forest (RF), and random tree (RT), for appraising the stability of a cohesive soil slope
Four conditioning factors that affected the values of factor of safety (FOS) were chosen: undrained shear strength (Cu), slope angle (β), setback distance ratio (b/B), and applied surcharge on the shallow foundation installed over the slope (w)

Summary

Introduction

The stability of local slopes is a critical issue that needs to be meticulously investigated due to their great impact on the adjacent engineering projects (excavation and transmission roads, for instance). Chakraborty and Goswami estimated the FOS for 200 slopes with different geometric and shear strength parameters using the multiple linear regression (MLR) and ANN models. They compared the obtained results with an FEM model, and an acceptable rate of accuracy was obtained for both applied models. In this paper, we investigated the efficiency of seven machine-learning models (i.e., in their optimal structure), namely multiple linear regression (MLR), multi-layer perceptron (MLP), radial basis function regression (RBFR), improved support vector machine using sequential minimal optimization algorithm (SMO-SVM), lazy k-nearest neighbor (IBK), random forest (RF), and random tree (RT), for appraising the stability of a cohesive soil slope. The design solution charts were developed using the outputs of the most efficient model

Data Collection and Methodology

Machine Learning Techniques

7: Support

Results and Discussion

Network Results

Conclusions