Metaheuristic-hybridized MLS-SVR: a multi-output model for predicting continuous shear displacements of rock fractures

Abstract

This study presents a metaheuristic-hybridized model based on sparrow search algorithm (SSA) and multi-output least-squares support vector regression machines (SSA-MLS-SVR) to predict the continuous shear displacements of rock fractures, which is closely related to the geo-structure stability and safety. To validate the reliability and potential of the proposed model, which was respectively developed by using two subsets of MDST database including 362 results of direct shear tests for rock fractures from laboratory and field. For the unsatisfactory generalization of preliminary model, three kinds of nonlinear transformations were utilized in data preprocessing to improve the data sensitivity of SSA-MLS-SVR. The performance of modified model indicated that the SSA-MLS-SVR can effectively grasp the correlation among each post-peak shear displacements in the continuous shear process of rock fractures. In addition, at the end of this paper, some interesting findings and conjectures about the potential connections among continuous shear displacements will be summarized. This study has a great significance of exploring the correlation among the continuous shear displacements of rock fractures, and the use of the proposed data processing method is valuable for further improving the model prediction performance.