【作者】 李国良； 李忠富；

【Author】 LI Guo-liang,LI Zhong-fu (School of Management,Harbin Institute of Technology,Harbin 150001,China)

【机构】 哈尔滨工业大学管理学院；

【摘要】 针对现有文献在进行企业绩效评价时,大多数模型与方法都没有考虑某些指标中存在的异常值将会对指标权重及评价结果产生影响的现状,本文以中国建筑业上市公司绩效评价为例,对评价中指标权重的计算过程进行了改进。首先通过聚类分析将那些具有较大相似性的评价对象划分成组,然后将组视为新的评价对象,以信息熵所反映的数据本身的效用值来计算指标的权重系数,最后再应用所得到的权重对中国2007年33家建筑业上市公司的绩效状况进行综合评价。研究表明,这种改进在一定程度上降低了异常值对于评价结果客观性的影响。更多还原

【Abstract】 The existence of abnormal values of certain indicators would impact on the weight of the indicators and impact on the assessment results in the corporate performance assessment.But the most of previous methods have not considered the situation.In order to reduce the effects,the weight calculation process was improved in the performance assessment of listed companies in China’s construction industry in this paper.First,the objects which are similar to each other are divided into groups through cluster analysis.And then,it taks the groups as new objects to calculate the weight coefficient of indicators through the information entropy that the data reflected.Finally,evaluates the performance of the 33 listed companies in China’s construction industry in 2007 with the calculated weight.The results show that the improvement reduce the impact of abnormal values on the assessment to a certain extent.更多还原