在对数据包分析(DEA)模型发展历程回顾的基础上,为了提高投入、产出指标选取以及权重确定的客观性,解决投入、产出指标存在隐变量以及测量指标间存在关联性或因果关系等问题,引入结构议程模型(SEM)约束锥,提出带有SEM约束锥的DEA模型.该模型利用SEM对原因变量和结果变量进行回归分析,客观确定指标权重,进而利用SEM方法对DEA模型权重的选择加以限制.利用该模型对中国区域建筑产业竞争效率进行实证分析结果表明,通过该模型缩小了分析结果与实际状况之间的偏差,同时提高了分析的精度.
In order to improve the objectivity of selecting input, output indicators and determining their weights, based on systematic review of the development of DEA model, SEM constraint cone and a DEA model bound with SEM restraint cone were introduced in this paper, which also contributed to solve problems such as existence of hidden variables in the input and output indicators and measures of correlation or causal relation among indicators. The DEA model first used SEM to analyze independent and dependent variables, determining weights of each indicator. Then the selecting of weights in DEA model was restricted by means of SEM. In the end, an empirical research on the validity of construction industry competitiveness of Chinese provinces was done by using this model, and the results show that the model reduces the deviation between the analytical results and the actual situation, while improving the accuracy of the analysis.