Intelligent Prediction and Parameter Optimization Analysis of Shield Construction Attitude Based on Random Forest Model

Abstract

Aiming at the problem that the attitude deviation is affected by multi-parameter coupling and the prediction accuracy is insufficient in the process of shield construction, this paper uses the random forest algorithm to construct the attitude deviation prediction model based on the field measured construction data. By sorting and optimizing the importance of input parameters, eliminating redundant features and optimizing the model structure, the accurate prediction of attitude deviation is realized. The results show that the goodness of fit (R²) of the model after parameter optimization is 0.917, the error indexes are significantly reduced, and the predicted values are in good agreement with the real values. The total thrust, cutterhead torque and earth pressure are the key parameters affecting the attitude deviation. The research shows that the random forest model based on parameter optimization can effectively improve the accuracy of attitude prediction, which can provide technical reference for the intelligent control of shield construction alignment.