Early Warning Model for Tunnel Construction Risks Based on Back Propagation Neural Network

In order to warn the safety risk in the process of tunnel construction, based on the neural network and the safety risk characteristics of tunnel construction, a tunnel safety risk warning model is established, and the risk of tunnel construction accidents are effectively warned. Taking the current situation of a tunnel construction as an example, the neural network model is trained and applied by using the expert scoring results, and the applicability of the model is verified. By introducing the SHAP (Shapley Additive Explanations) methodology, the interpretability of the model is enhanced, making the risk warning results more transparent and credible.