突发事件应急血液需求预测研究

发布时间：2018-03-23 02:23

本文选题：应急血液　切入点：需求预测　出处：《上海交通大学》2013年硕士论文　论文类型：学位论文

【摘要】：近年来，各种自然灾害和公共卫生突发事件频发，严重影响着人们日常生活和生命财产安全。血液是应急救援的重要物资，属于易腐品并具有很高的时间敏感性。准确预测突发事件应急血液需求对于保障血液供应，防止采血过量，避免灾后血荒等具有重要意义。但由于突发事件血液需求存在较强的突发性和波动性，与突发事件等级、受灾地区人口密度等有关，且在灾后不同救援阶段表现出不同的需求特点，给预测带来了很大难度，在目前的研究中尚缺乏有效的方法和手段。因此，本文对突发事件应急血液需求预测问题进行了系统的研究。本文在深入分析应急血液需求特征的基础上，提出针对紧急救援阶段和后续救援阶段的血液需求预测模型。首先，根据紧急救援阶段血液需求激增且呈现出明显的非线性、血液需求影响因素众多的特点，，本文提出了基于小波分析和支持向量机的组合预测方法，并采用Libsvm软件对模型进行求解。汶川地震案例的研究表明，该方法针对短期血液需求问题具有较高的预测精度。其次，在后续救援阶段，血液需求主要受次生衍生等二次灾害的影响，且存在伤员在不同救援区域间的转移。针对该阶段血液需求随机波动的特点，本文建立了加权马尔可夫链预测模型。汶川地震实际案例的仿真结果表明，本文针对后续救援阶段所提出的长期预测模型比传统的预测模型具有更高的准确性和适用性。本文的研究成果可以丰富目前应急血液需求领域的定量化研究，此外，该研究完善了应急救援体系，能够为血液救援工作提供科学的决策依据。
[Abstract]:In recent years, a variety of natural disasters and public health emergencies occur frequently, seriously affecting people's daily life and life and property safety. Blood is an important material for emergency relief. Perishable and highly time-sensitive. Accurate prediction of emergency blood demand for emergencies can protect blood supply and prevent excessive blood collection. The prevention of post-disaster blood shortage is of great significance. However, because of the strong sudden and volatile demand for blood in emergencies, it has something to do with the level of emergencies and the population density in the affected areas. And in different rescue stages after the disaster, it shows different demand characteristics, which brings great difficulty to the prediction, and lacks effective methods and means in the current research. In this paper, the prediction of emergency blood demand is studied systematically. Based on the in-depth analysis of the characteristics of emergency blood demand, this paper puts forward a prediction model of blood demand for emergency rescue stage and follow-up rescue stage. Firstly, according to the surge of blood demand in emergency rescue phase, it shows obvious nonlinearity. There are many factors affecting blood demand. In this paper, a combined prediction method based on wavelet analysis and support vector machine is proposed, and the model is solved by Libsvm software. This method has high prediction accuracy for short-term blood demand problem. Secondly, in the follow-up rescue stage, the blood demand is mainly affected by secondary disasters such as secondary derivation, and there exists the transfer of the wounded between different rescue areas. In this paper, a weighted Markov chain prediction model is established. The simulation results of Wenchuan earthquake case show that the long-term prediction model proposed in this paper is more accurate and applicable than the traditional prediction model. The research results of this paper can enrich the quantitative research in the field of emergency blood demand at present. In addition, the research improves the emergency rescue system and provides scientific decision basis for blood rescue work.
【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：R459.7

【参考文献】