主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

Chinese Journal of Management Science ›› 2011, Vol. 19 ›› Issue (3): 182-192.

Previous Articles    

The Ecosystem Evaluation Model Based on Kernel Principle Component Analysis and Its Applications

CHENG Yan-qiu, CHI Guo-tai   

  1. School of Management, Management, Dalian University of Technology, Dalian 116024, China
  • Received:2010-05-10 Revised:2011-03-26 Online:2011-06-30 Published:2011-06-30

Abstract: This paper constructes the ecosystem comprehensive evaluation index system framework using Driving-force-Pressure-State-Impact-Response(DPSIR)conceptual model. The evaluation model is established using kernel principle component analysis(KPCA)to deter mine the index weight and improved normalization method to compute index score. At the same time,the paper makes empirical analysis on the tenvice-province cities. The contributions lie on four aspects. Firstly,the paper constructes index system using DPSIR conceptual framework,and bring human dem and,economic development and social progress into the system. It is much better than the existing research which only considered the status of environment and resources or the status of environmental protection and pollution treatment. Secondly,the paper determines the optimal kernel function and the optimal kernel parameters by comparing the similarity of the best kernel matrix and the kernel matrix of the alternative kernel function. This solves the problem of choosing the kernel function and kernel parameters. Thirdly,the paper analyzes the key factors of ecological condition through Rclustering. This changes the conditions that the meaning of principle component is not clear and the existing researches are unable to determine the key factors. Fourthly,the result reflects that the major factors affecting the ecological condition are the percapita green GDP,the percapita green area,the air quality and the SO2 emission intensity of perten thousand yuan GDP. Where as the percapit a electricity consumption of resident,population density and the natural growth rate of population have less effects1 Viewing from criteria layer,the importance of criteria layers can be ordered as followed:Driving- force,State,Pressure and Response.

Key words: ecosystem evaluation, DP SIR conceptual framework, kernel principle component analysis, clustering analysis

CLC Number: