Research on prediction method of surrounding rock deformation of highway tunnel based on optimized Bayesian algorithm

In order to solve the problems of low prediction accuracy and poor generalization ability of the current tunnel surrounding rock deformation, this paper proposes a method based on Bayesian optimization long short-term memory network (LSTM). This method first analyzes the vault settlement Preprocess the original monitoring data of the highway tunnel vault settlement and peripheral convergence, and then construct an initial LSTM model of the highway tunnel vault settlement and peripheral convergence, and use Bayes to optimize the hyperparameters in the model to finally obtain the prediction results.