Load-settlement response of a footing over buried conduit in a sloping terrain: a numerical experiment-based artificial intelligent approach

Settlement estimation of a footing located over buried conduit in sloping terrain is challenging task for practicing civil/geotechnical engineers. In the recent past, advent machine learning technology has made many traditional approaches antiquated. This paper investigates viability, development, implementation, and comprehensive comparison five artificial intelligence-based models, namely multilayer perceptron, Gaussian processes regression, lazy K-Star, decision table, random forest (RF) to estimate settlement within soil slope. The pertaining dataset 3600 observations was obtained by conducting large-scale numerical simulations via finite element modeling framework. After executing feature selection technique that correlation-based subset selection, applied load, total unit weight soil, constrained modulus slope angle ratio, hoop stiffness conduit, bending burial depth crest distance were utilized as influence parameters estimating forecasting settlement. predictive strength accuracy all models mentioned supra evaluated using several well-established statistical indices such Pearson’s correlation coefficient (r), root mean square error (RMSE), Nash–Sutcliffe efficiency (NSE), scatter index (SI), relative percentage difference (RPD). results showed among employed this study, perceptron model shown better with r, RMSE, NSE, SI, RPD values (0.977, 0.298, 0.937, 0.31, 4.31) (0.974, 0.323, 0.928, 0.44, 3.75) training testing dataset, respectively. sensitivity analysis revealed selected play an important role determining output value. However, modulus, weight, have highest importance 18.4%, 16.3%, 15.3%, 13.8%, 11.4%, Finally, translated into functional relationship easy implementation can prove useful practitioners researchers predicting terrain.