Two key challenges for reliability analysis of slopes considering spatial variability of geotechnical parameters are high dimensionality and computational cost. Therefore, this study proposes a slice inverse regression (SIR) based extreme gradient boosting (XGBoost) method for reliability analysis of spatially variable slopes, which combines the advantages of both SIR and XGBoost. The Karhunen-Loève (K-L) expansion is adopted for spatially variable soil property modeling. Then, the slope reliability analysis based on the SIR-XGBoost method is proposed. In this study, the proposed method is illustrated using two typical slopes of single and double layers as examples. The results show that the reliability analysis based on SIR-XGBoost can predict the failure probability of slopes with reliable accuracy and efficiency. Compared with the results obtained by other methods, the stability and generalization of the proposed method in computational efficiency are further expounded.

Reliability evaluation of slope in spatially variable soils using sliced inverse regression‑based extreme gradient boosting