DOC: stats.boxcox_normmax: correct minimize -> maximize

[mdhaber]

Reference issue

Closes gh-18748

What does this implement/fix?

gh-18748 noted that scipy.stats.boxcox_normmax documentation states

Minimizes the log-likelihood boxcox_llf.

but boxcox_llf is the (positive) log-likelihood function to be maximized, not minimized. This PR corrects the mistake.

Maximization of the log-likelihood function is accomplished by minimizing the negative log-likelihood:

scipy/scipy/stats/_morestats.py

Lines 1256 to 1259 in 2e16ec5

	def _mle(x):
	def _eval_mle(lmb, data):
	# function to minimize
	return -boxcox_llf(lmb, data)

This PR clarifies this point in the documentation of the optimizer parameter.

[tupui]

Thanks Matt, LGTM. Good rewriting and clearer with respect to [negative]log-likelihood