Study On Emergency Blood Demand Forecasting

With the development of modernization and industrialization,environmental degradation has been increasingly serious and the livingsystem of human beings has been much more vulnerable. Human beingsare suffering from various kinds of emergency events, including naturaldisasters and public healthy emergencies. Scholars have focused onemergency management in order to develop the abilities of respond andreduce the losses. Emergency resources demand forecasting plays a keyrole as it acts as the premise and basis of optimal allocation of emergencyresources. Blood is a kind of special emergency resource because it ishighly time sensitive for its function as life support. Blood is also scarceand excess blood collection is not commonly accepted because it maylead to social problems, such as blood supply shortage.The research on emergency blood demand forecasting is conducted.This dissertation firstly summarizes the related theories involvingemergency management, demand forecasting methods and emergencyresources demand forecasting. Based on analyzing the characteristics ofemergency blood demand, this dissertation proposes the theoreticalanalysis framework for the forecasting issue. The analysis highlights theimportance of dividing the whole rescue process into two parts, and eachproper forecasting method is applied into each one. Two forecastingmethods are proposed which includes the combined model composed ofWavelet Transform (WT) and Support Vector Machine (SVM) to meetdemands for the crucial rescue period, and the Markov model to meet thelong-term demands during the given periods. The numerical results ofWenchuan earthquake happened in China prove both methods proposedare effective.