The combination prediction methods of mixed model used in subsidence monitoring
Qiu We · Engineering of Surveying and Mapping · 2014
In order to predominate the regular pattern,predicte the deform trend and guarantee the security of the building,it is necessary to carry on modeling forecast with the data of building settlement.Every prediction model has its own applicable situation and shortcoming which is unable to meet current accuracy requirements.Two commonly-used prediction models,gray GM(1,1)and autoregressive model are selected in two different ways.The subway settlement monitoring data of Baijiahu station in Nanjing Metro Line one are calculated and analyzed which shows that the two combinations have an improved forecast accuracy compared with single model.