Is there a documentation, paper etc. where it is explained why scikit-learn does not provide p-values/confidence levels (1, 2, 3, 4)? Note: I'm not asking about opinions, but about documentation. For example the R package lme4 does not provide p-values either, but has a FAQ point about this. On the official scikit-learn website I found no explanation for this desicion.
$\begingroup$
$\endgroup$
$\begingroup$
$\endgroup$
There is a scoring function in sklearn package which is used in a feature selection procedure. This function first computes F-score and then converts it to p-value.
answered Jan 27 '20 at 10:03
$\begingroup$
$\endgroup$
Several feature selection techniques return p values, e.g. ANOVA, chi2,..:
https://scikit-learn.org/stable/modules/generated/sklearn.feature_selection.f_classif.html
