Prediction of Marshall Stability and Marshall Flow of Asphalt Pavements Using Supervised Machine Learning Algorithms

The conventional method for determining the Marshall Stability (MS) and Flow (MF) of asphalt pavements entails laborious, time-consuming, expensive laboratory procedures. In order to develop new advanced prediction models MS MF current study applied three soft computing techniques: Artificial Neural Network (ANN), Adaptive Neuro-Fuzzy Inference System (ANFIS), Multi Expression Programming (MEP). A comprehensive database 343 data points was established both MF. nine most significant straightforwardly determinable geotechnical factors were chosen as predictor variables. root squared error (RSE), Nash–Sutcliffe efficiency (NSE), mean absolute (MAE), square (RMSE), relative (RRMSE), coefficient determination (R2), correlation (R), all used evaluate performance models. sensitivity analysis (SA) revealed rising input significance results parametric (PA) also found be consistent with previous research findings. findings comparison showed that ANN, ANFIS, MEP are reliable effective methods estimation mathematical expressions derived from represent novelty relatively simple. Roverall values in > ANFIS ANN over permissible range 0.80 Therefore, techniques higher performance, possessed high generalization capabilities, assessed parameters terms training, testing, validation sets their closeness ideal fit, i.e., slope 1:1, outperformed other two this will contribute choice an appropriate artificial intelligence strategy quickly precisely estimate Parameters. Hence, would assist safer, faster, more sustainable predictions MF, standpoint time resources required perform tests.