Fuzzy Bayesian Classification of LR Fuzzy Numbers

Abstract

Fuzzy data is considered as an imprecise type of data with a source of uncertainty. Fuzzy numbers allow us to model uncertainty of data in an easy way which justifies the increasing interest on theoretical and practical aspects of fuzzy arithmetic. This paper presents a Fuzzy Bayesian Classifier (FBC) over LR-type fuzzy numbers with unknown conditional probability density function. A new version of K-NN method is used to estimate conditional probability density function for Bayesian classification of fuzzy numbers. Fairly good recognition rate has been obtained over fuzzy numbers in classification using FBC even in the presence of noise.