I would like to implement a BUGS/WINBUGS or JAGS model that specifies the likelihood node by marginalizing over two parameters. Is this feasible in BUGS/WINBUGS or JAGS?

This is the posterior \begin{align} &P\left( g_{x} , g_{y} | \mathbf{a}\right) \propto P\left( \mathbf{a} | g_{x} , g_{y}\right) P\left( g_{x} , g_{y}\right) \end{align} and here is the likelihood \begin{align} &P\left( \mathbf{a} | g_{x} , g_{y}\right) = \int_{\boldsymbol{\omega}_{R}}\int_{\boldsymbol{\omega}_{T}}P\left( \mathbf{a} | g_{x} , g_{y} , \boldsymbol{\omega}_{R} , \boldsymbol{\omega}_{T}\right)P\left( \boldsymbol{\omega}_{R}\right) P\left( \boldsymbol{\omega}_{T}\right)\mathrm{d}\boldsymbol{\omega}_{R} \mathrm{d}\boldsymbol{\omega}_{T} \end{align} where \begin{align} \boldsymbol{\omega}_{R} \\ \boldsymbol{\omega}_{T} \\ g_{x}\\ g_{y} \\ \end{align} are all stochastic nodes


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.