BriefGPT.xyz
Apr, 2023
变量复杂度加权温和吉布斯采样器用于贝叶斯变量选择
Variable-Complexity Weighted-Tempered Gibbs Samplers for Bayesian Variable Selection
HTML
PDF
Lan V. Truong
TL;DR
本文提出了一种新的分组加权温度调整 Gibbs 采样器无需计算后验包含概率 (PIP) 以降低总的计算量,我们还提出了与之相关的 Rao-Blackwellized 估计器的方差上界,并通过实验表明,相对于现有的子集 wTGS 方法,我们的方法具有更小的方差。
Abstract
subset weighted-tempered gibbs sampler
(wTGS) has been recently introduced by Jankowiak to reduce the computation complexity per MCMC iteration in high-dimensional applications where the exact calculation of the
posteri
→