Doc

Generalized Extreme Value Distribution

docs/tutorial:stats:continuous_genextreme

Extreme value distributions with one shape parameter $c$ .

If $c > 0$ , the support is $- \infty < x \leq 1/ c .$ If $c < 0$ , the support is $\frac{1}{c} \leq x < \infty.$

μ_{n}^{'} = \frac{1}{c ^{n}} k = 0 \sum n (k n) (- 1)^{k} Γ (c k + 1) if c n > - 1

So,

For $c = 0$ the distribution is the same as the (left-skewed) Gumbel distribution, and the support is $R$ .

Implementation: scipy.stats.genextreme