中文摘要：

评价指标体系建立的是否合理决定着人们能否对评价对象有个正确的认识。本文提出一种基于改进灰色关联度的评价指标体系构建方法，该方法首先计算出各因素与系统的关联度，采用德尔菲法对专家打分进行处理，打分结果作为各因素重要性的得分；然后将因素的关联度与重要性得分相结合，排序后筛选出初始指标；最后通过对初选指标之间的相关性分析，求解出有效的指标体系。比较实验是在真实的数据集上进行的，实验结果证明改进的灰色关联度分析法明显优于主成分分析法。因此可以认为改进后的方法能够有效的实现指标筛选。

英文摘要：

Whether evaluation objects can be understood comprehensively or not depends on a reasonable evaluation index system. In this paper, a novel evaluation index system construction method based on improved grey correlative degree analysis has been proposed. Firstly, the correlation between various factors and the system has been calculated, dealing with expert scores based on the Delphi method and assigning the results to the importance score of each factor. Secondly, the correlative degree has been combined with the importance score and the initial indicators have been screened after sorting. Finally, an effective index system has been drawn out through analyzing the correlations of primary indexes. Comparison experiment has been conducted on real data sets. The experimental results demonstrate that the improved grey correlation analysis method outperforms the principal component analysis method. It suggests that the proposed method can be applied for achieving effective index screening.