期刊文献+

线性科技评价中自然权重问题及修正研究——动态最大均值逼近标准化方法 被引量:12

Studying and Correction on the Problem of Natural Weights in Linear Science and Technology Evaluation:The Normalization Method of Approximate Dynamic Maximum Mean
下载PDF
导出
摘要 由于评价指标不服从正态分布、评价指标均值不相等,导致即使标准化后的评价指标,也存在隐含的自然权重问题,会造成评价实际权重的严重扭曲,并影响评价结果。将权重分为设计权重、自然权重、实际权重,从理论上分析了三者之间的关系,并提出了动态最大均值逼近标准化方法,彻底消除了自然权重问题。以JCR 2015经济学期刊评价为例,说明了自然权重对评价的不良影响,在消除自然权重以后,使得实际权重与设计权重完全一致,保证了评价的客观性与公正性。该方法也适用于反向指标,只不过首先对该指标要进行正向化处理。 Because the evaluation indexes do not obey the normal distribution,and the average of the evaluation indexes is unequal.So even after the standardization of evaluation index,there are implicit natural weight problems.It can cause serious distortion of the actual weights on the evaluation,and affect the evaluation results.In this paper,the weights are divided into design weights,natural weights and actual weights.The relationship between the three kinds of weights is analyzed theoretically,and the standardization method of approximate dynamic maximum mean is proposed,which eliminates the natural weight problem completely.Taking the JCR 2015 Journal of economics as an example,which explain the adverse effects of natural weights on evaluation.After the natural weights are eliminated,the actual weights are consistent with the design weights,thus ensuring the objectivity and impartiality of the evaluation.The method is also applicable to the reverse index,but need handling the indexes positively firstly.
作者 俞立平 YU Li-ping(School of Management and E-businessl,Zhejiang Gongshang University,Hangzhou 310018,China)
出处 《统计与信息论坛》 CSSCI 北大核心 2018年第10期27-33,共7页 Journal of Statistics and Information
基金 教育部人文社会科学基金项目<协同创新深度的影响机制与对策研究>(17YJA630125) 国家自然科学基金面上项目<基于子群视角的多指标群组集成技术研究>(71671165) 浙江省哲学社会科学规划项目<经济群组评价技术及应用研究>(17NDJC211YB)
关键词 自然权重 设计权重 实际权重 动态最大均值逼近标准化 natural weights design weights actual weights standardization method of approximate dynamic maximum mean
  • 相关文献

参考文献16

二级参考文献120

共引文献269

同被引文献152

引证文献12

二级引证文献29

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部