Sorry, I’m having difficulty expressing myself clearly. I want to find out the equation that is generating the values for a given model. In simple models, such as a non-hierarchical linear regression, the likelihood function is obvious (just the univariate normal pdf). However for my model, which is complex and hierarchical, I can build it in pymc3 easily but don’t know how to write the likelihood symbolically. Presumably there is such a symbolic expression of the likelihood/posterior within Theano somewhere.
1 Like