arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2303.08653math.STecon.EMecon.THstat.TH
收藏

关于后验均值的稳健性

On the robustness of posterior means

Jiafeng Chen

AI总结:

该研究针对正态位置模型,在两类均值为零、方差有界的先验下,证明了满足尾部条件时后验均值的平方误差贝叶斯风险具有一致有界性,揭示了后验均值的稳健性。

摘要:

考虑方差σ²已知的正态位置模型X∣θ~N(θ,σ²)。假设θ服从先验分布G₀,该先验均值为零且方差有上界V。令G₁为可能存在误设的先验,同样均值为零且方差有上界V。我们证明,在G₁满足额外尾部条件的前提下,G₁下后验均值的平方误差贝叶斯风险是有界的,且该界对所有G₀、G₁和σ²>0一致成立。

英文摘要:

Consider a normal location model $X \mid θ\sim N(θ, σ^2)$ with known $σ^2$. Suppose $θ\sim G_0$, where the prior $G_0$ has zero mean and variance bounded by $V$. Let $G_1$ be a possibly misspecified prior with zero mean and variance bounded by $V$. We show that the squared error Bayes risk of the posterior mean under $G_1$ is bounded, subjected to an additional tail condition on $G_1$, uniformly over $G_0, G_1, σ^2 > 0$.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑