Fisher Neyman factorisation theorem

469 Views Asked by At

What's the best place to read a proof of the full-generality Fisher Neyman factorisation theorem? I have a few stats books that claim to give a proof but they leave out steps like, for instance, why a dominated family of distributions is equivalent to the mixture of a countable subfamily -- and they're just generally irritating for other reasons.

1

There are 1 best solutions below

2
On

Not a bad question. A paper by Halmos and Savage claimed to do this, and I heard there was a gap in the argument, consisting of a failure to prove certain sets have measure zero:

  • P. R. Halmos and L. J. Savage, "Application of the Radon–Nikodym theorem to the theory of sufficient statistics," Annals of Mathematical Statistics, volume 20, (1949), pages 225–241.

I've heard the the first fully sound proof was by R. R. Bahadur, but I'm not succeeding in finding that right now.

It's easy to prove in the discrete case.