derekbeaton / OuRS

Outliers and Robust Structures
GNU General Public License v3.0
11 stars 0 forks source link

score correlations #13

Open derekbeaton opened 5 years ago

derekbeaton commented 5 years ago

I'm not so sure those are correct anymore... because of arbitrary flips.

derekbeaton commented 5 years ago

(1) I think I know the solution, and (2) the correlations should change over to abs() or cor()^2 regardless... the direction is irrelevant (and actually problematic).

derekbeaton commented 4 years ago

because I'm terrible at comments, this is w.r.t. split-half/two-fold repeated approaches