Paper Info
Title
Importance weighting and andness control in De Morgan dual power means and OWA operators
Abstract
Importance weighted aggregation plays a central role in utilization of information resources for information retrieval and fusion, pattern and object recognition, decision making, etc. A class of aggregation operators of particular interest is formed by the aggregation operators between the min (minimum) and the max (maximum), the so-called averaging operators. Two key issues in the choice of such an operator for a given application are the kind of importance weighting and the andness (''minness'') of the operator. Two main kinds of importance weighting for such operators, namely multiplicative and implicative, are proposed and discussed. The purpose of this paper is to facilitate the choice and application of such operators through providing a systematization of their classes according to their behavior and equipping some classical averaging operators, namely the power means and the OWA operators, with importance weighting schemes and direct parametric andness control for both kinds of importance weighting. For increased efficacy and for symmetric behavior by andness and orness (=1-andness) at the same degree of both measures, the two classes of averaging operators are applied in a De Morgan dual form. The main issue in the choice of underlying the classical averaging operator appears to be the computational cost of its application.
Year
DOI
Venue
2012
10.1016/j.fss.2011.04.001
Fuzzy Sets and Systems
Keywords
Field
DocType
importance weighted aggregation,information resource,implicative importance weighting,aggregation operator,importance weighting scheme,orness,andness,importance weighting,so-called averaging operator,averaging operators,owa operator,information retrieval,classical averaging operator,de morgan dual power,de morgan dual averaging operators,aggregation operators,direct parametric andness control
Discrete mathematics,Mathematical optimization,Weighting,Multiplicative function,Parametric statistics,Operator (computer programming),Mathematics,Cognitive neuroscience of visual object recognition
Journal
Volume
ISSN
Citations 
196,
Fuzzy Sets and Systems
6
PageRank 
References 
Authors
0.51
17
1
Name
Order
Citations
PageRank
Henrik Legind Larsen154545.16