Application of CPSO-LSSVM model based on wavelet transform and Kalman filter to the subsiding analysis of structure
Hongwei Gao, Hongyan Wen, Ruibo Fang, Guang-Yu Nie, Zhi Xiang Yang · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2015
In the analysis on building deformation, when the deformation is unconscious, the significant effect random noise contributes, in addition that the prediction model of neural network has slow convergence. A new CPSO-LSSVM forecasting model is established, based on the combination of wavelet analysis and Kalman filter. The forecasting model resolves the noise problem and with a high accuracy by improving the PSO algorithm. The results show that it has a higher accuracy than the BP network and the LSSVM.