A maximum-entropy approach to the linear credibility formula

► We employ the MEM to approximate a Bayes' estimator with a linear credibility formula. ► We establish the approximation coincides with the credibility under some conditions. ► We provide an increasing weight for more recent observations.