Study on combined prediction model for number of transferred and resettled people in rainstorm-flood disaster
In order to predict the number of people who need to be transferred and resettled under the rainstorm-flood disas-ters more scientifically and accurately,the cases of severe rainstorm-floods in China from 2011 to 2018 were collected,and the relationship between the number of transferred and resettled people and the influencing factors representing the severity of rainstorm-flood disaster was tested by Pearson correlation analysis.Then,the regression model based on principal component analysis(PCA)and the support vector machines(SVM)were used to predict the number of people who need to be trans-ferred and resettled under rainstorm-flood disaster.Based on the results of the two methods,a combined prediction method was proposed to revise the number of transferred and resettled people under rainstorm-flood disaster.The results show that both the MSE and MAE of the combined prediction method are less than those of regression prediction and SVM model prediction.Using the combined prediction method to predict the number of transferred and resettled people in flood disaster can fully combine the advantages of single prediction model and improve the prediction accuracy and generalization ability of the com-bined prediction model.The research results can provide a reference for determining the sheltering needs of rainstorm-flood disasters and formulating the sheltering evacuation plans.
rainstorm-flood disasternumber of transferred and resettled peoplecombined predictionsupport vector ma-chines(SVM)
NETL
NSTL
万方数据
