基于主成分-熵的评价指标体系信息贡献模型

科研管理 ›› 2014, Vol. 35 ›› Issue (12): 137-144.

基于主成分-熵的评价指标体系信息贡献模型

迟国泰¹, 李战江^1,2

1. 大连理工大学工商管理学院, 辽宁大连 116024;
2. 内蒙古农业大学理学院, 内蒙古呼和浩特 010018

收稿日期:2013-06-09 修回日期:2014-03-28 出版日期:2014-12-25 发布日期:2014-12-23
作者简介:迟国泰(1955-),男(汉),黑龙江海伦人,大连理工大学工商管理学院教授、博导,研究方向:复杂系统评价、风险管理。
李战江(1977-),男(汉),内蒙古乌海人,管理科学与工程专业博士生,内蒙古农业大学讲师,研究方向:复杂系统评价、风险管理。
基金资助:
国家自然科学基金项目(71171031,2012.01-2015.12);教育部科学技术研究项目(2011-10,2011.07-2014.10);国家自然科学基金青年科学基金项目(71201018,2012.07-2016.07);中国银监会银行业信息科技风险管理项目(2012-4-005,2012.02-2013.10);大连银行小企业信用风险评级系统与贷款定价项目(2012-01,2012.01-2013.12)。

Model of information contribution of evaluation index system based on principal component- entropy

Chi Guotai¹, Li Zhanjiang^1,2

1. School of Business Management, Dalian University of Technology, Dalian 116024, Liaoning, China;
2. College of Science, Inner Mongolia Agricultural University, Hohhot 010018, Inner Mongolia, China

Received:2013-06-09 Revised:2014-03-28 Online:2014-12-25 Published:2014-12-23

摘要/Abstract

摘要： 评价指标体系的建立是各种评价问题研究中的不可缺少的重要研究内容,如果最终建立的评价指标体系的信息量损失过大,则无论建立什么样的评价指标体系都没有意义。通过将主成分分析方法与信息熵方法相结合,本文建立了最终构建的评价指标体系相对于海选评价指标体系的信息贡献测算模型。本文的创新与特色一是通过使用主成分方法与信息熵方法度量指标体系的信息含量,构建了评价指标体系相对于海选指标体系的信息贡献测算模型,改变了现有研究中只考虑评价指标如何筛选而忽视筛选出的评价指标相对于海选指标的信息贡献,解决了评价指标体系相对于海选指标体系的信息贡献的测算问题。二是实例计算表明从海选指标中筛选出的评价指标保留了94.4%的海选指标信息。

关键词: 评价指标体系, 海选指标体系, 信息贡献, 主成分, 熵

Abstract: The establishment of evaluation index system in various evaluation researches is an important and indispensable research content. If the information content of the eventual evaluation index system has too much loss, any evaluation index system makes no sense. Combining principal component analysis method and information entropy method, this paper establishes the measurement model of the information contribution of eventual evaluation index system compared to the mass-election evaluation index system. In this paper,the main characteristics and innovations lie in the following aspects. First, it uses principal component analysis and information entropy to measure information contents of the index system, to establish the measuring model of information contributions of the evaluation index system compared to the mass-election index system, to change the existing studies which only consider the selection problem of the evaluation index and neglect the information contribution problem, and to solve the measuring problem of the information contributions. Second, the example shows that eventual evaluation index system retains 94.4% information content of mass-election evaluation index system.

Key words: evaluation index system, mass-selection index system, information contribution, principal component, entropy

中图分类号:

N945.16

迟国泰, 李战江. 基于主成分-熵的评价指标体系信息贡献模型[J]. 科研管理, 2014, 35(12): 137-144.

Chi Guotai, Li Zhanjiang. Model of information contribution of evaluation index system based on principal component- entropy[J]. Science Research Management, 2014, 35(12): 137-144.

参考文献

[1] Cristian SC, Monica D. Entropic measures, markov information sources and complexity [J]. Applied Mathematics and Computation, 2002, (132): 369-384. [2] Ali D, Ehsan SS, Refik S. Information measures for generalized gamma family[J]. Journal of Econometrics, 2007, (138): 568-585. [3] Hu BG, Wang Y. Evaluation criteria based on mutual information for classifications including rejected class[J]. Acta Automatica Sinica, 2008, 34(11): 1396-1403. [4] Xu WH, Zhang XY, Zhang WX. Knowledge granulation, knowledge entropy and knowledge uncertainty measure in ordered information systems[J]. Applied Soft Computing, 2009, (9): 1244-1251. [5] Zhang QS, Jiang SY, Jia BG,et al. Some information measures for interval-valued intuitionistic fuzzy sets [J]. Information Sciences, 2010, (180): 5130-5145. [6] Yann R. A Radon-Nikodym approach to measure information [J]. Mathematical Social Sciences, 2011, (61): 170-177. [7] Dai J, Xu Q.Approximations and uncertainty measures in incomplete information systems[J]. Information Sciences, 2012, 198: 62-80. [8] Yu S, Huang TZ, Liu X, et al. Information measures based on fractional calculus[J]. Information Processing Letters, 2012, 112(23): 916-921. [9] 方先明. 信息量与预测模型优劣的评判[J]. 统计与决策, 2006(16): 158-159. [10] 张登兵. 基于事实比较和效用的信息测度方法研究[J]. 中国集体经济, 2008(24): 77-78. [11] 胡丹, 李洪兴, 余先川. 规则与规则库信息量的度量及其应用[J]. 中国科学F辑: 信息科学, 2009, 39(2): 218-233. [12] 左淑霞, 席建锋, 肖殿良等."特色交通标志"设计及其信息量度量方法研究[J]. 中国安全科学学报, 2010, 20(11): 139-144. [13] 顾雪松, 迟国泰, 程鹤. 基于聚类-因子分析的科技评价指标体系构建[J]. 科学学研究, 2010, 28(4): 508-514. [14] 王晓勇, 楼佩煌, 唐敦兵. 基于信息量的不确定型多属性决策方法[J]. 运筹与管理, 2012, 21(1): 64-69. [15] 高惠璇. 应用多元统计分析[M]. 北京: 北京大学出版社, 2005: 16-290. [16] 杨波, 苏娜. 基于熵理论的高科技企业投融资绩效研究—以湖北省为例[J]. 科研管理, 2011, 32(8): 42-50. [17] 林正奎. 基于熵权-AHP组合的城市保险业社会责任评价研究[J]. 科研管理, 2012, 33(3): 142-160. [18] 张良, 蔡生. 信息量的度量及应用[J]. 沈阳大学学报, 2004, 16(2): 89-91.