大坝变形的XGBoost-LSTM变权组合预测模型及应用

XGBoost-LSTM Combinatorial Model with Variable Weight for Dam Deformation Prediction and Its Application

为了实现更高精度的大坝变形预报,提出了一种大坝变形的XGBoost-LSTM变权组合预测模型,即首先引入XGBoost模型和LSTM模型对大坝变形分别进行分析预测,然后采用变权组合方法将二者的分析预测结果进行有机融合,进而得到最终预测结果。以某混凝土重力坝为例,首先通过与随机森林、ELMAN以及逐步回归分析各模型的对比研究,论证了XGBoost与LSTM应用于大坝变形预测的优越性;进一步地,XGBoost与LSTM的变权组合预测效果相较于各单一模型取得了较大程度的提升,且相较于二者的等值赋权组合提升优势更加显著,变形预测结果与工程实际情况更加吻合,具有较好的适用性和可推广价值。

A XGBoost-LSTM combinatorial model with variable weight is proposed to more accurately predict dam deformation.First,the XGBoost(eXtreme Gradient Boosting)model and LSTM(Long Short-Term Memory)model are introduced to analyze and predict the dam deformation respectively,and then the results of the two models are combined by using variable weight combination method to obtain the final prediction result.With a concrete gravity dam as a case study,the advantages of XGBoost and LSTM models in dam deformation prediction are demonstrated respectively through comparison with those of random forest,ELMAN and stepwise regression analysis models;furthermore,the prediction effect of the combinatorial model is verified to have enhanced remarkably compared with each of the single model and the equivalent-weighted XGBoost-LSTM combinatarial model.The deformation prediction results are more consistent with the actual engineering situation,thus is well applicable and popularizable.