An asymmetry coefficient for random variables based on fuzzification

In this paper we consider a new asymmetry coefficient for random variables. This measure of skewness is based on a fuzzy representation of real-valued random variables which characterize the distribution of the original random variable through the expected value of the 'fuzzified' one. This representation has been recently introduced by the authors in a statistical inferential context. On the basis of a simple random sample from a random variable, a point estimator of the skewness of the variable is presented, and some relevant properties of this estimator are studied.