I'm a beginner in stats and I have a very basis question on GLM.
Suppose Gaussian distribution. GLM tries to do this:
$$g(E(y|X))=X\beta$$
where $g$ is a link function.
But if I transform $y$ before regression:
$$y'=g(y)$$
and do this:
$$E(y'|X)=X\beta$$
now this is an ordinary linear regression.
So my question is how are these two operations different with each other?
