Prediction Reinforced Slope Stability Using Pile Using Adaptive Neuro-Fuzzy Inference System (ANFIS) Model

Predictive analysis using artificial intelligence (AI) has transformed the landscape of forecasting analysis in various research fields. The advancements in AI modelling algorithms have enhanced decision-making, trend identification, and process optimization. In geotechnical engineering, AI assists in predicting soil behaviour, structural stability, and slope stability. The AI model discussed in this paper is the Adaptive Neuro-Fuzzy Inference System (ANFIS). In this study, the ANFIS model predicts slope stability by examining the Factor of Safety (FOS) value. Slope stability analyses reinforced with continuous bored pile walls generated by the numerical computation of the finite element method (FEM) in two dimensions (2D) and three dimensions (3D) are compared with the predictions of the ANFIS model. The numerical FEM computations employ PLAXIS 2D and PLAXIS 3D software. Meanwhile, the ANFIS model is designed within the MATLAB software platform involving 112 data samples. With six input pile parameters and one output, the finding shows that the ANFIS model can learn complex non-linear data and accurately predict the output. This is supported by the R² values of 0.9771 and 0.9965 from comparing the forecasting output with the 2D and 3D FEM outputs, respectively. Meanwhile, the low RMSE values of 0.0187 and 0.0180 each confirm this.