Prediction of geomechanical parameters using soft computing and multiple regression approach

The evaluation of geotechnical parameters of geo-materials are essential part of every geotechnical project. But sometimes, it is not possible to determine the all required parameter in the laboratory. Therefore, scientist and engineers used the statistical and empirical relation to determine the crucial parameters. The present study focused on the determination of parameters like uniaxial compressive strength (UCS), tensile strength (TS), point load index (PLI) and Young's modulus (YM) from very easily determinable physical parameters viz, density (DEN), porosity (PORO) and compressional wave velocity (P-WV) using multiple variable regression analysis (MVRA) and adaptive neuro-fuzzy inference system (ANFIS). The various ANFIS structures and MVRA models were tried for prediction of desired parameters and best one was considered based on variance account for (VAF), root mean square error (RMSE) and correlation coefficient (R-2). ANFIS structure not only depends on the input parameters and rules, but also on the output parameter as observed in case of PLI. (C) 2016 Elsevier Ltd. All rights reserved.