What is the Box-Cox transformation?

medium

Answer

$\mathrm{Box} - \mathrm{Cox}(y, \;{\lambda})\; = \;(y{\lambda}\; - \;1)\; / \;{\lambda}$
$Choose λ (typically via MLE) to make the transformed variable closest to normal.$
$Extended to negative values by Yeo-Johnson.$
$Uses: normalize skewed regression targets, improve linear-model residual normality.$
$In modern ML, tree models don't need it; useful for linear / GLM diagnostics.$

Check yourself — multiple choice

Random transformation
$(y{\lambda}\; - \;1) / {\lambda}$ (log at λ=0); pick λ via MLE to maximize normality; Yeo-Johnson handles negatives
Only for classification
Same as log

Box-Cox: parametric power transform; λ maximizes normality; Yeo-Johnson for negatives.

#descriptive

215 interview questions in this topic.