Rock Strength Estimation Using Several Tree-Based ML Techniques

The uniaxial compressive strength (UCS) of rock is an essential property material in different relevant applications, such as slope, tunnel construction, and foundation. It takes enormous time effort to obtain the UCS values directly laboratory. Accordingly, indirect determination through conducting several index tests that are easy fast carry out interest importance. This study presents powerful boosting trees evaluation framework, i.e., adaptive machine, extreme gradient machine (XGBoost), category for estimating sandstone. Schmidt hammer rebound number, P-wave velocity, point load were chosen considered factors forecast sandstone samples. Taylor diagrams five regression metrics, including coefficient (R2), root mean square error, absolute variance account for, A-20 index, used evaluate compare performance these trees. results showed proposed able provide a high level prediction capacity prepared database. In particular, it was worth noting XGBoost best model predict achieved 0.999 training R2 0.958 testing R2. had more outstanding capability than neural network with optimization techniques during phases. performed variable importance analysis reveals has significant influence on predicting