I know a way to visualize outliers is to make a box plot, but wanted to know if scikit had any quick ways to detect outliers for each variable?
-
$\begingroup$ Do you want a univariate approach, or multivariate? $\endgroup$ – Jon Nordby Feb 13 '20 at 8:50
-
$\begingroup$ @jonnor univariate is fine $\endgroup$ – Insu Q Feb 13 '20 at 12:10
The scikit-learn implementations of Isolation Forest and Local Outlier Factor should help you out here. Details on both at this link:
https://scikit-learn.org/stable/modules/outlier_detection.html