A new definition of cross-entropy for uncertain variables

Cross-entropy for uncertain variables is used to measure the divergence between two uncertainty distributions. Logarithm cross-entropy and quadratic cross-entropy for uncertain variables fail to measure the degree of divergence associated with some uncertain variables; thus, this paper proposes a new definition of cross-entropy for uncertain variables as a supplement and discusses its properties. A formula of cross-entropy is derived via inverse uncertainty distributions. Moreover, this paper also defines the generalized cross-entropy of uncertain variables and investigates its properties.