Title
Integration of correlations with standard deviations for determining attribute weights in multiple attribute decision making
Abstract
This paper proposes a correlation coefficient (CC) and standard deviation (SD) integrated approach for determining the weights of attributes in multiple attribute decision making (MADM) and a global sensitivity analysis to the weights determined. The CCSD integrated approach determines the weights of attributes by considering SD of each attribute and their CCs with the overall assessment of decision alternatives, where CCs are determined by removing each attribute from the overall assessment of decision alternatives. If the CC for an attribute turns out to be very high, then the removal of this attribute has little effect on decision making; otherwise, the attribute should be given an important weight. The global sensitivity analysis to the weights of attributes is proposed to ensure the stability of the best decision alternative or alternative ranking. A numerical example about the economic benefit assessment of the industrial economy of China is investigated to illustrate the potential applications of the CCSD method in determining the weights of attributes. Comparisons with existing weight generation methods are also discussed.
Year
DOI
Venue
2010
10.1016/j.mcm.2009.07.016
Mathematical and Computer Modelling
Keywords
Field
DocType
multiple attribute decision,overall assessment,correlation coefficient,important weight,decision alternative,standard deviation,attribute weights,attribute weight,global sensitivity analysis,multiple attribute decision making,alternative ranking,best decision alternative,economic benefit assessment,ccsd integrated approach,ccsd method,rank correlation
Correlation coefficient,Data mining,Ranking,Variable and attribute,Correlation,Global sensitivity analysis,Statistics,Standard deviation,Numerical stability,Mathematics
Journal
Volume
Issue
ISSN
51
1-2
Mathematical and Computer Modelling
Citations 
PageRank 
References 
63
2.01
16
Authors
2
Name
Order
Citations
PageRank
Ying-Ming Wang13256166.96
Ying Luo246720.48