A machine learning regression approach for predicting the bearing capacity of a strip footing on rock mass under inclined and eccentric load

In this study, the Multivariate Adaptive Regression Splines (MARS) model is employed to create a data-driven prediction for bearing capacity of strip footing on rock mass subjected an inclined and eccentric load. The strengths masses are based Hoek-Brown failure criterion. To develop set training data in MARS, lower upper bound finite element limit analysis (FELA) carried out obtain numerical results with width B . There six considered dimensionless variables, including geological strength index ( GSI ), constant/yield parameter m i γB/σ ci adhesion factor α load angle from vertical axis β eccentricity e/B ). A total 5,120 FELA solutions P/σ ) obtained used as set. influences all variables factors mechanisms investigated discussed detail. sensitivity these also examined.