Intelligently Predict the Rock Joint Shear Strength Using the Support Vector Regression and Firefly Algorithm

Abstract To propose an effective and reasonable excavation plan for rock joints to control the overall stability of surrounding mass predict prevent engineering disasters, this study is aimed at predicting joint shear strength using combined algorithm by support vector regression (SVR) firefly (FA). The dataset collected was employed as output prediction, roughness coefficient (JRC), uniaxial compressive (σc), normal stress (σn), basic friction angle (φb) input machine learning. Based on database strength, training subset test learning processes are developed realize prediction evaluation processes. results showed that FA can adjust hyperparameters effectively accurately, obtaining optimized SVR model complete strength. For testing results, able obtain values 0.9825 0.2334 determination root-mean-square error, showing good applicability SVR-FA establish nonlinear relationship between variables Results importance scores σn most important factor affects while σc has least significant effect. As a influencing stiffness from perspective physical appearance, change JRC value impact