基于PCA-BBO-SVM的尾矿坝变形预测模型与性能验证研究*

为准确预测尾矿坝变形趋势，通过主成分分析法（PCA）对尾矿坝变形影响因子进行优选，基于生物地理学优化算法（BBO）对支持向量机（SVM）参数进行寻优，建立PCA-BBO-SVM尾矿坝变形预测模型，并以杨家湾尾矿坝为例对模型性能进行验证。研究结果表明:PCA-BBO-SVM模型在4个测点的RMSE为0.139 6，0.274 2，0.317 0，0.530 6；MAE为0.112 5，0.213 5，0.269 0，0.412 9；MAPE为0.525 0%，0.692 3%，2.621 2%，1.311 2%；预测精度及对局部波动的预测能力均高于BP、GS-SVM、GA-SVM和PSO-SVM模型，研究结果可为尾矿坝变形预测提供模型支撑。

Prediction model of tailings dam deformation based on PCA-BBO-SVM and its performance verification

In order to accurately predict the deformation trend of tailings dam,the principal component analysis method (PCA) was used to optimally screen out the influencing factors of tailings dam deformation,and the parameters of support vector machine (SVM) were optimized based on the biogeographic optimization algorithm (BBO).A PCA-BBO-SVM prediction model of tailings dam deformation was established,and the Yangjiawan tailings dam was taken as an example to verify the performance of the model.The results showed that the RMSE of the PCA-BBO-SVM model at four measuring points were 0.139 6,0.274 2,0.317 0 and 0.530 6,the MAE were 0.112 5,0.213 5,0.269 0 and 0.412 9,and the MAPE were 0.525 0%,0.692 3%,2.621 2% and 1.311 2%,respectively.The prediction accuracy and the ability to predict local fluctuation were higher than those of BP,GS-SVM,GA-SVM and PSO-SVM models,and it can provide technical support for the deformation prediction of the tailings dam.