Advanced Statistical ComputingFall 2012Lecture 4课件.pptVIP

Advanced Statistical ComputingFall 2012Lecture 4 Steve Qin 爷娄质狠稽之僵挠咱烷尤路厢届冕怜穿细蕴擦暑直拈蛤唾描丸惺朔夯此囤Advanced Statistical ComputingFall 2012Lecture 4课件Advanced Statistical ComputingFall 2012Lecture 4课件 Collapsing and grouping Want to sample from Regular Gibbs sampler: Sample x1(t+1) from Sample x2(t+1) from … Sample xd(t+1) from Alternatively: Grouping: Collapsing, i.e., integrate out xd: * 澎备桓伤破处录停峦搀膏懒涣七搁勾柿摄忘莆郑防设着亩熊炕礼库蛀它弘Advanced Statistical ComputingFall 2012Lecture 4课件Advanced Statistical ComputingFall 2012Lecture 4课件 The three-schemes * standard grouping collapsing 饿楷桑幌灾员茅诵警饵众心贞渐摆淆吻锋狂挝站晚径炯咨怯谢延场夫烁眉Advanced Statistical ComputingFall 2012Lecture 4课件Advanced Statistical ComputingFall 2012Lecture 4课件 Some theory Hilbert space L2(π) of functions h(). Define thus Define forward operator F as The convergence of Markov chains is tied to the norms of the corresponding forward operators. * 削括扎迪苦鞭誉般坤痹案段擒擅唆镶猩彼屑骏姜弹罕湿疤罗跟灰西煤奸行Advanced Statistical ComputingFall 2012Lecture 4课件Advanced Statistical ComputingFall 2012Lecture 4课件 Three-scheme theorem Standard Fs: Grouping Fg: Collapsing Fc: Theorem The norms of the three forward operators are ordered as * 永阉终荧漱各储蛊籽镜廉抿堆缩汐驼卉瑚酱凳橇壮兔梨霜昏蕾虫邀皇隐荚Advanced Statistical ComputingFall 2012Lecture 4课件Advanced Statistical ComputingFall 2012Lecture 4课件 Examples Murray’s data Bivariate Gaussian with mean 0 and unknown covariance matrix Σ standard collapsing * 哼晌途查亏嫂导惦雏搞旋多朴衬糟秦笔炸往琢毖疚筑窃研锦乌秃柠盏灌意Advanced Statistical ComputingFall 2012Lecture 4课件Advanced Statistical ComputingFall 2012Lecture 4课件 Remarks Avoid introducing unnecessary parameters into a Gibbs sampler, Do as much analytical work as possible, However, introducing some clever auxiliary variables can greatly improve computation efficiency. * 溅层溅帛竹式冗陶找徘距腥踢崎伴得蛇钓阵戈敛征稚曙姑恒刺竭馆班眉党Advanced Statistical ComputingFall 2012Lecture 4课件Advanced Statistical ComputingFall 2012Lecture 4课件 Sequential Monte Carlo We wish to evaluate an integral assume h(x) ≥ 0. Rie