The Markov moment problem and de Finetti's theorem. I (Q1762690)

[For part II of this paper see the following review, Zbl 1066.60005.] This work has to do with the Markov moment problem: Given a positive real number \(c\) and a sequence \(s_0, s_1, \ldots\) of real numbers what are necessary and sufficient conditions for the existence of a probability measure \(\mu\) on \([0,1]\) such that (i) \(\{s_n\}\) is the moment sequence of \(\mu,\) (ii) \(\mu\) is absolutely continuous, and (iii) \(d\mu/dx\) is almost everywhere bounded above by \(c.\) The absolute continuity above is defined relative to Lebesgue measure. In this work this last measure is replaced by any other probability measure \(\nu\) on \([0,1].\) Also, characterizations of \(L_p\) densities for \(p>1\) are considered instead of \(L_{\infty}\) density above. Then the connection to de Finetti's theorem is done with characterizations of the mixing measures. Some applications to Bayesian statistics are presented. Monotone densities are also characterized.