考虑邻近点相关性的EN-SVM组合混凝土双曲拱坝变形预测模型

模型的输入通常决定模型的性能,而在普遍的混凝土坝变形预测模型中往往仅考虑以环境因素作为输入情况.基于此,在考虑环境因素影响的基础上,把邻近测点的相关性也纳入考虑,构建了考虑邻近点相关性的弹性网络(EN)与支持向量机(SVM)的组合预测模型,即先构造两类不同的特征,然后分别使用主成分分析法(PCA)和EN进行类内特征处理...

EN-SVM Combined Deformation Prediction Model for A Concrete Hyperbolic Arch Dam Considering the Correlation of Adjacent Points

The input of the model usually determines the performance of the model, but in the general prediction model of concrete dam deformation, only environmental factors are considered. On the basis of considering the influence of environmental factors and taking the correlation of adjacent measuring points into account, the prediction model of a concrete hyperbolic arch dam combined with elastic network(EN) and support vector machine(SVM) was constructed considering the correlation of adjacent points. Firstly, two kinds of different characteristics were constructed. Then, principal component analysis(PCA) and EN were used to process the in-class features, and the processed feature factors were input into the EN and SVM models respectively to obtain the predicted values of each model. Finally, the optimal weight recombination method was used to combine the predicted results of the two models to obtain the final predicted values. The results of engineering examples show that the root mean square error, mean absolute error and mean absolute percentage error of the combined model are lower than those of other models, which verifies its effectiveness and superiority.