Nonparametric Sequential Bayes Estimation of the Distribution Function

摘要: Abstract The paper considers nonparametric estimation of the distribution function under weighted squared error loss plus cost. Sequential Bayes rules are proposed Dirichlet process priors. Since implementation procedures via backward induction (cf. Arrow et al., 1949) is computationally prohibitive, we consider in this approximations to sequential rules. Like other problems type, it turns out, case, that asymptotically pointwise optimal (APO) stopping Bickel and Yahav, 1967 1968 1969a b) equivalent myopic or one-step look-ahead Also, provide a proof first-order optimality second-order expansion APO Recommended by Michael Baron