The Dirichlet distribution and process through neutralities (Q2641428)

Some new characterizations are given for the Dirichlet distribution and Dirichlet process in terms of neutrality and neutrality to the right, e.g., Theorem 8: if \((F(t))_{t\in\mathbb R}\) is a stochastic process such that its trajectories are a.s. CDFs, \(( F(t))_{t\in\mathbb R}\) is neutral to the right and \(1-F(t_n)\) is neutral in \((F(t_1), F(t_2)-F(t_1),\dots, F(t_n)-F(t_{n-1}), 1-F(t_n))\) for any \(n\geq 2\) and any \(t_1<\dots<t_n\), then \(( F(t))_{t\in\mathbb R}\) is a Dirichlet process.