On the edited fuzzy K-nearest neighbor rule

摘要

Classification of objects is an important area in a variety ofnfields and applications. In the presence of full knowledge of thenunderlying joint distributions, Bayes analysis yields an optimalndecision procedure and produces optimal error rates. Many differentnmethods are available to make a decision in those cases whereninformation of the underlying joint distributions is not presented. Thenk nearest neighbor rule (k-NNR) is a well-known nonparametric decisionnprocedure. Many classification rules based on the k-NNR have alreadynbeen proposed and applied in diverse substantive areas. The edited k-NNRnproposed by D.L. Wilson (1972) would be an important one. Fuzzy theory,noriginated by L.A. Zadeh (1965), is widely used to represent thenuncertainty of class membership. The fuzzy k-NNR has been proposed bynseveral investigators. In this paper an edited type of the fuzzy k-NNRnis developed. Next, some asymptotic properties of the proposed editednfuzzy k-NNR are created. Moreover, numerical comparisons are madenbetween the proposed edited fuzzy k-NNR and the other fuzzy k-NNR. Thosenresults confirm that the edited fuzzy k-NNR has a better performancenthan the fuzzy k-NNR