Premium
Density‐induced ordered weighted averaging operators
Author(s) -
Ma FengMei,
Guo YaJun
Publication year - 2011
Publication title -
international journal of intelligent systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.291
H-Index - 87
eISSN - 1098-111X
pISSN - 0884-8173
DOI - 10.1002/int.20500
Subject(s) - operator (biology) , mathematics , measure (data warehouse) , constraint (computer aided design) , similarity (geometry) , argument (complex analysis) , degree (music) , dispersion (optics) , mathematical optimization , algorithm , computer science , artificial intelligence , data mining , physics , biochemistry , chemistry , geometry , repressor , transcription factor , acoustics , optics , image (mathematics) , gene
Abstract We provide a special type of induced ordered weighted averaging (OWA) operator called density‐induced OWA (DIOWA) operator, which takes the density around the arguments as the inducing variables to reorder the arguments. The density around the argument, which can measure the degree of similarity between the argument and its nearest neighbors, is associated with both the number of its nearest neighbors and its weighted average distance to these neighbors. To determine the DIOWA weights, we redefine the orness measure, and propose a new maximum orness model under a dispersion constraint. The DIOWA weights generated by the traditional maximum orness model depend upon the order of the arguments and the dispersion degree. Differently, the DIOWA weights generated by the new maximum orness model also depend upon the specific values of the density around the arguments. Finally, we illustrate how the DIOWA operator is used in the decision making, and prove the effectiveness of the DIOWA operator through comparing the DIOWA operator with other operators, i.e., the centered OWA operator, the Olympic OWA operator, the majority additive‐OWA (MA‐OWA) operator, and the kNN‐DOWA operator. © 2011 Wiley Periodicals, Inc.