Moment representations for generalized Lehmer polynomials

Source abstract

We represent normalized generalized Lehmer polynomials as moments of a random variable constructed from two independent negative binomial variables and an independent uniform variable. An integral form of their known exponential generating function yields mixtures of uniform distributions and convergent double series. We recover Eulerian polynomial identities and give a probabilistic interpretation of anti-diagonal sums of poly-Bernoulli numbers. We also prove moment determinacy, characterize the Stieltjes parameter range, derive Hankel total non-negativity and log-convexity, and identify the boundary distributions.