嵌套并联结构两阶段DEA下科技创新效率测度与分解

目前有许多研究将两阶段DEA模型应用到科技创新活动当中，一般将创新活动分为研发和转化两个阶段，为了进一步探讨不同转移转化方式下科技成果的转化效率以获得更多的详细信息，本文拓展了传统的两阶段DEA模型，考虑了嵌套并联结构的两阶段网络DEA模型，将科技创新活动分为科技成果研发阶段和科技成果转化阶段，并将科技成果转化阶段分为内部转化和外部转化两个并联子系统。本文选取14家中科院院属单位作为被评价对象，应用加权加法分解方法计算整体效率和各阶段子效率。通过将本文模型与传统两阶段模型进行对比发现，传统模型会低估科技成果转化效率。结果表明：整体科技创新效率均值偏低，科技成果研发效率均值高于科技成果转化效率均值，科技成果转化效率仍有很大的提升空间。科技成果转化阶段中内部转化效率普遍较高但是外部转化效率普遍较低，较低的外部转化效率是导致科技成果转化效率低下的主要原因。最后，本文将所有被评价单位分为四类机构，分析提高科技创新效率的实现路径，可以通过重点突破模式和渐进突破模式来实现整体效率的提高。

Measuring S&T Efficiency and its Decomposition using a Two-stage DEA Model with Nested Parallel Structure

Existing studies have applied the two-stage DEA model to the innovation activities and divided it into two stages:R&D process and transformation process.In order to obtain more information about different modes of transformation efficiency, the traditional two-stage DEA model is extended with the parallel structure inside,which is comprised of internal and external transformation. 14 institutes of the Chinese Academy of Sciences are selected as DMUs, and the weighted addition decomposition method is applied to calculate the overall efficiency and the efficiency of each stage.By contrasting to the traditional two-stage DEA model,the former model underestimates the transformation efficiency. The results show:The mean of total S&T innovation is at a low level and the R&D process efficiency is higher than the transformation efficiency,so it still needs to improve the efficiency of transformation.In the second process,internal transformation efficiency is higher than external, which contributes to the low transformation efficiency.Finally, all DMUs are divided into four categories and the realization path about how to improve S&Tinnovation is analyzed, by proposing the key breakthrough mode and the gradual breakthrough mode.