Emergency Material Demand Prediction Based On Case Consumption Reasoning

The response and treatment of emergency is a research field which has been developed late and developed slowly in China.Among them,earthquake disasters,as one of the most serious natural disasters,are a long-standing problem in dealing with emergency events.Because of the uniqueness of China's geographic location,earthquake disasters occur frequently.In view of the great necessity of earthquake emergency rescue,the prediction of emergency material requirements has an irreplaceable position and importance for emergency rescue and post-disaster reconstruction.In the current study of emergency material demand forecasting,case-based reasoning is more suitable for the current data,among which there are few analyses on attribute weight selection and few models to analyze model robustness under different weights.In the case-based reasoning study,there are few studies on the characteristics of data,the time is a important one in emergency data attribute study,and few previous studies have considered time attribute.In order to obtain a reasonable plan of material supply in the early stage,according to the characteristics of emergency data,the factors that are easy to obtain in the early stage of the earthquake are selected,a basic theory of case consumption is given,and a direct prediction method of case reasoning is used.With this as the core,an emergency material demand prediction model is built to study the problem of emergency material demand.First,in terms of attribute selection,in order to adapt to the urgency of earthquake rescue,the selection of attributes that are easy to obtain at the initial stage is more in line with the actual situation;secondly,in view of the high dimensionality of earthquake emergency data and the prone to existence of multiple secondary attributes,a relatively low-level attribute is introduced.A new weight calculation method specifically for emergency data-the rough set attribute dependence enhancement method,which assigns attribute weights after data preprocessing;considering the real-time nature of emergency material demand forecasting and the time characteristics of emergency data,this article adopts For the understanding of metabolic thinking,a new idea is proposed to weaken the weight of the similarity of old cases when calculating the similarity of case-based reasoning.With this idea as the core,the case consumption theory is given and the case consumption reasoning model is constructed;finally,considering the method of this paper Robustness and the degree of influence of the weight calculation method on the effect of the case consumption strategy.This article gives two weight calculation methods commonly used in emergency material demand forecasting,namely,the entropy weight method and the coefficient of variation method.Further prove the effectiveness of the case consumption strategy.Experiments have proved that when using case-based reasoning to predict emergency supplies,the average relative error between the material prediction results using the consumption strategy and the actual results is 0.354 lower than that of the traditional method;an important change parameter-consumption interval is given in the consumption strategy Parameter,this parameter determines the strength of the consumption capacity.Experiments have found that the consumption interval parameter has a great influence on the overall prediction effect.The reasonable choice of the consumption interval parameter can optimize the prediction result of the case consumption reasoning;the final robustness Experiments show that the change of the weight method does not affect the optimization results of the case consumption reasoning.The other two weight selection methods are better than the traditional case reasoning method when using the consumption strategy,but the average relative error of the rough set attribute dependence enhancement method is higher.Low,more suitable for the case consumption reasoning of the data set in this article.Through innovative research on emergency material demand forecasting,it is possible to quickly predict the supply of materials at the early stage of the earthquake and provide a basis for the decision-making plans of relevant departments.