Post-Earthquake Material Dynamic Demand Forecast Based on XGBoost
As a sudden natural disaster,earthquake often causes serious casualties,and the research on emergency material de-mand prediction is an important part of post-disaster emergency rescue,and it is the premise of overall material allocation.This pa-per divides the prediction of the demand for emergency supplies after the earthquake into two steps:firstly,the total number of casu-alties of the earthquake is estimated,and then the empirical function is used to estimate the initial number of earthquake casualties,and the purpose is to achieve a more accurate prediction of the subsequent casualties according to the casualties in the previous peri-od of time,and the XGBoost-based seismic dynamic casualties prediction model is introduced based on the casualties in the previous time to improve the accuracy and reliability of the prediction.Secondly,the linear relationship between the material demand and the number of dynamic casualties is used to calculate the daily material demand according to the dynamic demand estimation model with the concept of lead time,so as to realize the dynamic demand forecasting.Finally,the proposed method is used to dynamically pre-dict the number of casualties,injuries and deaths of the"Wenchuan earthquake",and the dynamic demand for materials is esti-mated,which provides a reference for the supply of emergency materials in the disaster area.
earthquake disastercasualty forecastXGBoostlead timedemand estimation model
万方数据
维普
