Study on Seismic Damage Prediction Model of Urban Bridges Group Based on PSO-SVM

Ertao Wang,Huiying Gao,Hai Sun,Junjie Wang
DOI: https://doi.org/10.11899/zzfy20170119
2017-01-01
Abstract:According to the observed urban bridge damage data,the particle swarm optimization (PSO) was used to optimize the input parameters of support vector machine (SVM) model.Eight factors that effect bridge seismic damage level are chosen as the input vector.By making full use of the advantages of PSO and SVM,we establish the PSO-SVM model.By comparing the urban bridge damage prediction ability of the SVM model and PSO-SVM model,we conclude that the PSO-SVM model has relatively high accuracy and strong generalization capability,which is of important reference and guide value.
What problem does this paper attempt to address?