变量降维及零值数据在DEA应用中的改进

Variables Reduction and Zero Data in Data Envelopment Analysis

随着DEA方法的不断发展,实际应用中发现了不少影响DEA效率评价的因素。其中,决策单元自由度的限制及评价指标出现零值数据是本文在对油田区块勘探效率评价时所遇到的问题。提出一种改进的变量合成方法对评价指标进行降维,对零值的数据进行合成。将合成后的数据代入CCR模型及超效率DEA模型进行验证发现:改进后的变量指标数据提高了DEA模型有效性,且超效率DEA模型的评价结果更精确。

The DEA method has been used widely in many aspects of our life,while there are a number of pitfalls that could have a major influence on the efficiency.In this paper,we have found that the freedom of degree associated with the DMU and zero data are the two important problems in evaluating the exploration efficiency of oil block.Therefore,we present an improved method of variable reduction,which can solve the problem of zero data in DEA.Then,we verify the CCR and super-efficiency DEA model with the above processed data,and find that the method can improve the efficiency of DEA model.Moreover,the results of super-efficiency DEA are more accurate.