实验室评估指标有效性的实证研究

An empirical study on the validity of laboratory evaluation indexes

摘要:

现阶段国家、省市高度重视科研绩效评估工作,对实验室开展绩效评估可以有效反映实验室建设情况和科研水平,有助于提升实验室建设质量和高效管理。文章通过无监督学习和经典分类算法进行绩效指标评估体系进行整体分析,并引入专家评分结果进行实证研究,发现绩效评估指标体系整体具有可分性和有效性,但在实际运用中存在主、客观评估差异,提出采用信息熵计算与决策树相结合的方法进行指标体系的筛选和引导权重设置,并通过实验证明该方法具有较好的分类性能,能够更好地分析指标的有效性。

Abstract:

At present, the state and provinces attach great importance to the performance evaluation of scientific research, which can effectively reflect the laboratory and scientific research level, and help to improve the quality and efficient management of laboratory . Through unsupervised learning and classical classification algorithm, this paper makes an overall analysis of the performance index evaluation system, and introduces the results of expert evaluation for empirical research. It is found that the overall performance evaluation index system is separable and effective, but there are subjective and objective evaluation differences in the actual application. It is proposed to use the method of information entropy calculation and decision tree to screen the index system Through experiments, it is proved that this method has better classification performance and can better analyze the effectiveness of indicators.