Discriminant Analysis for Fuzzy Random Variables Based on Nonparametric Regression

This communication is concerned with the problem of supervised classification of fuzzy data obtained from a random experiment. The data generation process is modelled through fuzzy random variables which, from a formal point of view, can be identified with a kind of functional random element. We propose to adapt one of the most versatile discriminant approaches in the context of functional data analysis to the specific case we handle. The discriminant analysis is based on the kernel estimation of the nonparametric regression. The results are applied to an experiment concerning fuzzy perceptions and linguistic labels Keywords— fuzzy data, random experiments, supervised classification, kernel estimation, nonparametric regression.