Tunnel geomechanical parameters prediction using Gaussian process regression

The purpose of this study is to apply a modern intelligent method of Gaussian process regression (GPR) to predict the geological parameter of Rock Quality Designation (RQD) along the tunnel route. This method can also be used for any geological parameter prediction of tunnel future levels. The GPR method has been studied based on data obtained from 51 tunnels all over the world. Fifty data sets were utilized for intelligent modeling, while one of the data sets that belonged to Hamru tunnel in Iran, was used to evaluate the prediction approach. The comparisons’ results indicate that the GPR model’s prediction results are generally in good agreement with the actual results. The proposed GPR, on the whole, performs better than the support vector machine (SVM), artificial neural network (ANN) and linear regression (LR) in predictive analysis of the RQD parameter. •Tunnel geology parameters prediction model using Gaussian process regression (GPR).•Access to 50 learning data sets through the old tunnels all over the world.•Comparing the GPR results of with the results of SVR, ANN, DT, and LR.•Investigation of the prediction accuracy of each of the prediction models.