[WP34s] Regularized incomplete Beta function

[Paul Dale]

(05-08-2014 06:26 PM)Dieter Wrote:  Which leads to the question if an upper limit in the dof value would make sense. For instance a maximum of 1000. This would apply to the Student as well as the Chi² and F distribution.

Given we're only promising accuracy for single precision, there are many orders of magnitude left before the risk of digit loss becomes significant. I don't think I'd want to impose a limit suitable for DP that restricts results in SP, likewise I don't want a computation that errors out in DP but not SP.

For Student's-T and Chi² distributions we could approximate with a normal for very high degrees of freedom I guess. I don't remember the asymptotic properties of the F distribution when one or both degrees of freedom get large.


- Pauli