Closed iago-lito closed 2 months ago
Thanks for catching this, seems like there's similar - but less erroneous imo - usages of $k$ where the docs previously reference x
, there's more consistency in the docs using x
for the parameter of functions characterizing a distribution.
Would you be willing to add the ones in the distribution::hypergeometric
in this PR as well?
Generally, I have no opposition to readings of statements in the manner that always uses $x$ in terms of a function parameter,
$$ K \sim \textrm{Pois}(\lambda) $$
$$ \implies \textrm{Pr}(K = x) = f_\lambda(x) $$
As it's also good for other 1D random variables, even if multi parameter,
$$ X \sim \textrm{HypGeom}(N,n,\vec{p}); |\vec{p}|_1 = K $$
$$ \implies \textrm{Pr}(X = x) = f_{N,n,\vec{p}}\ (x) $$
as well as continuous and to the same notation for CDF as well, which could be defined for float even on discrete distributions (of appropriate support).
Hello. I am not exactly sure I understand your request, but maybe this additional commit is what you expect?
Yes! Thanks for figuring it out, I now realize how I could have better demonstrated what I meant.
For some reason, these formulae used a mixture of
k
andx
for the same parameter. I haven't checked whether this problem is spread more widely within the docs though.