构建的指标体系是否具有合理性直接影响着后续评价结果的准确性。文章立足于高校科技创新过程与区域创新系统分析角度，通过指标定性与定量筛选方法相结合构建高校科技创新能力评指标体系。本文的结论，一是运用复相关系数分析剔除易被其他指标替代的指标，避免指标间信息重叠；通过灰色关联分析剔除对评价目标不重要的指标，保留对评价结果影响显著的指标；二是计算了灰色关联分析中的分辨系数值，提高了灰色关联度的准确度；三是以我国31个省级行政区为样本，最终构建包含27个指标的基于区域视角的高校科技创新能力评价指标体系。四是最终筛选出的指标集保留了90.7%的原始指标集信息。

The validity of index system directly affects the accuracy of the evaluation results. Based on the process of scientific innovation and regional innovation system, this paper establishes an original index system for scientific and technology innovation ability of university with qualitatively and quantitively screening methods. The conclusions lie on the following aspects. First, the multiple correlation method is used to delete the indicator which reflects repetitive information with other indicators. The gray correlation method is used to delete the indicator which is easily replaced by others, and it guarantees that the screened indicators have significant effect on results. Second, the distinguishing coefficient was calculated which improves the accuracy of gray correlation method. Third, the index system for the university scientific and technology innovation ability is established with 31 provinces, which contains 27 indicators. Fourth, the example shows that eventual inex system contains 90.7% information content of original index system.

N945.16/C91-0

无