Generalized Discrimination Analysis


Sikharulidze A.

11th WSEAS International Conference on Automatic Control, Modelling and Simulation, İstanbul, Turkey, 30 May - 01 June 2009, pp.312-313 identifier

  • Publication Type: Conference Paper / Full Text
  • City: İstanbul
  • Country: Turkey
  • Page Numbers: pp.312-313

Abstract

In the following work the new method of decision-making is elaborated - generalized discrimination analysis, which is generalization of classical discrimination analysis [1] for the cases where the activities are fuzzy and represented by their compatibility values. The characteristic - frequency, which was used in classic discrimination analysis is replaced by fuzzy characteristic - Most Typical Value (MTV) [2], such as Fuzzy Expected Value (FEV)[3-5], Weighted Fuzzy Expected Value (FWEV)[6] and Generalized Weighted Fuzzy Expected Value (GWFEV)[7-8]. Respectively the method of decision-making is generalized, while the main idea remains the same - positive discrimination represents the belief that the activity is more indicative of the decision, than any of the remaining decisions, and negative - vice-versa. The proposition of generalization in the case when MTV=FEV and fuzzy measure is sampling distribution is proved. An example of use of generalized discrimination analysis is demonstrated. The application of given method in Psychiatry are presented.