Re: Distribution for the Linear Combination of the Product of 2 Gaussians

The toy problem you gave is normal and I based my answer on that.
Convoluting it is "easy" (using characteristic functions, that is ;-)
). I believe the convolution stuff has a good balance between
generality and usefulness.

If you want to be really really general about it, you should consider
solving it using the definition (i.e. borel sets and company). Well
it's a joke, but if you're the really pure mathematician that I think
you are, the idea may appeal to you ;-)

Any advanced prob book will have that. For something more concrete, my
teacher said a book authored by a certain Chou is a real classic. I
believe most contemporary advanced prob book has it in their
bibliography.

hth,
jon

.