false positives in mathematical text

[allefeld]

Here are a few false positives I found when running proselint on a mathematical-statistical text. Some are based on a misinterpretation of technical terms, another on not being aware of embedded math. The source is Pandoc-style Markdown with embedded LaTeX math. Not sure whether these can be fixed, but hope this helps.

by discarding all terms in the negative log-likelihood function

jargon.misc 'in the negative' is jargon. Can you replace it with something more standard?

linear discriminant $t$

needless_variants.misc Needless variant. 'discriminating' is the preferred form.

$$
H(y | T) = \frac12 \log_2 (\pi e ~ m n),
$$
but the marginal entropy
$$
H(y) = \int - f_y(y) \log_2 f_y(y) ~ \ud y
$$

misc.but No paragraph should start with a 'But'.

 were directly estimated from 1,000,000 simulated $t$-tests

dates_times.dates When specifying a date range, write 'from X to Y'.

[Ubadub]

@suchow What do you think of a "math" flag that would suppress such warnings?