Does anyone know how to get Stata to estimate, using max likelihood,
the parameters of a distribution using a sample from that
distribution? I.e. NOT to regress a dependent variable on a set of
indep. variables, but simply to estimate parameters.
The problem is that the ml model in Stata seems to require a regression
equation to be input as part of the maximising program and the model
instruction.
Even though the Stata documentation calls the 'theta's' "equations",
does not necessarily imply that you need to have data in them. You can
have a set of "equations" that only consists of constant terms: These
are simply the parameters of your distribution. The "equations" of
Stata's ml procedure does not have to correspond to equations in the
mathematical sense: Add one equation (in the Stata sense) per
parameter, use the /sigma1 notation (see "notes on syntax" in the
online help) to define a parameter `sigma1' you can use in the
likelihood expression.
