This expands on #7 : I think we need to add some features to the base distributions to allow the user to change the support through a linear shift. For example, in dpmm problems it is necessary to change the support of the standard exponential to accommodate the fact that alpha0 has domain -1 < alpha0 < infty. This is currently implemented in polya.
This expands on #7 : I think we need to add some features to the base distributions to allow the user to change the support through a linear shift. For example, in dpmm problems it is necessary to change the support of the standard exponential to accommodate the fact that alpha0 has domain -1 < alpha0 < infty. This is currently implemented in polya.