Assessment of the Rock Elasticity Modulus Using Four Hybrid RF Models: A Combination of Data-Driven and Soft Techniques

The determination of the rock elasticity modulus (EM) is an indispensable key step for design engineering problems. Traditional experimental analysis can accurately measure EM, but it requires manpower and material resources, time consuming. EM estimation new rocks using former published empirical formulas also a possibility be attached high uncertainties. In this paper, four types metaheuristic optimization algorithms (MOA), named backtracking search algorithm (BSA), multi-verse optimizer (MVO), golden eagle (GEO) poor rich (PRO), were utilized to optimize random forest (RF) model predicting EM. A data-driven technology was used generate integrated database consisting 120 samples from literature. To verify predictive performance proposed models, five common machine-learning models one formula developed predict Four popular indices, including root-mean-square error (RMSE), mean absolute (MAE), coefficient (R2) Willmott’s index (WI), adopted evaluate all models. results showed that PRO-RF has obtained most satisfactory prediction accuracy. porosity (Pn) important variable based on sensitive analysis. This paper compares RF optimized by MOA prediction. It provides good example subsequent application soft techniques other parameter estimations.