Combined modeling method of PCA-WOA-LSSVM for dam deformation prediction model

The traditional dam deformation prediction model is difficult to reflect the complex nonlinear relationship between the dam deformation and its multidimensional influencing factors. The least square support vector machine (LSSVM) can effectively deal with the nonlinear problem, so a combined modeling method of PCA-WOA-LSSVM is proposed to construct the dam deformation prediction model. Firstly, principal component analysis (PCA) is used to reduce the influence factors of dam deformation, reduce the input dimension of the model and reduce the complexity of the model. At the same time, whale algorithm (WOA) is used to search the optimal parameter combination of LSSVM model to improve the prediction accuracy of the model. Taking a dam as an example, compared with the dam deformation prediction models established based on PCALSSVM, LSSVM and traditional multiple regression methods, the model constructed based on PCA-WOA-LSSVM combined modeling method has better fitting ability and prediction accuracy, and can be better applied in dam deformation safety monitoring.