Effectiveness Of Decision Making Units Based On The Bootstrap Method Research

Judgment Decision Making Units (Decision Making Units,DMU) the effectivenessof the problem is an important problem in data envelopment analysis (DEA), over theyears, many researchers proposed various judgment method, the effectiveness of thedecision-making unit but in the process of judgment Decision unit is valid, interactionbetween Decision Making Units tend to be ignored. In view of this, this article on thebasis of predecessors’ judgment decision-making unit effectiveness of decision makingunits interact, excavate potential effective decision making units.Basic CCR model based on data envelopment analysis (dea), taking the Bootstrapmethods of statistics in a limited sample of repeated sampling to get a large number ofsample data, thus to map sample points of the histogram, according to the histogrammethod, it is concluded that the probability density distribution function considering theinteraction between the decision-making unit, should avoid strong decision-makingunits of other decision-making effectiveness, the influence of detection by using themethod of relative entropy given the lack of strong decision making units before andafter the change of the probability density distribution, thus achieve the purpose ofidentifying potential effective decision making units.The empirical analysis part through two examples verify the effectiveness of theproposed method. Instances of15deputy provincial city resources efficiencymeasurement analysis of input and output data, chengdu, xian, shenzhen, changchun,Harbin and ningbo after six influential city, potentially effective to reflect, in guangzhouby relative entropy of the inspection found that for the rest of the eight cities ofguangzhou influence is stronger, illustrates the method to find the potential effectivedecision-making units, prove the effectiveness of the proposed method; Instance2to thecost of funds investment, cost, risk and income index efficiency calculation analysis,also found the latent effectiveness in the fund effectively, further verify the effectivenessof the proposed method.In complete thesis also encountered many problems in the process, such as the useof the Bootstrap sample when the sample set and choice; In how to estimate relativeentropy calculation process of medium density function, etc., through rigorous anddetailed theoretical analysis and a large number of repeated experiments, finally solvethe problem encountered in the process of paper, effectively completed the method innovation.