Earth Pressure Multipoint Prediction for EPS Shield Based on Multi-Model Ensemble

The tunnel face stability is greatly influenced by the earth pressure in the chamber. It is of great practical significance to improve the performance of the earth pressure multipoint predictive model. In this study, a multi-model ensemble approach based on Lasso, Support Vector Regression, Random Forest and Gradient Boosting Decision Tree is presented for earth pressure multipoint prediction in EPB shield. The Leave-One-Out is adopted to validate the predictive performance. The feature importance is provided by the Lasso, Random Forest and Gradient Boosting Decision Tree model. The experimental results show that the performance of the multi-model ensemble is better than all single model.