fleiss_kappa#

scikitplot.corpus.fleiss_kappa(ratings, *, labels=None)[source]#

Fleiss’ kappa for any number of raters with nominal labels.

Parameters:

ratingsarray-like of shape (n_items, n_raters)

One label per rater per item; every item has the same number of ratings (at least two) and none is missing.

labelssequence or None, default=None

The label set; None uses the labels seen.

Returns:

float

NaN (with a warning) when every rating is the same label.

Raises:

ValueError

For fewer than two raters, a missing value or an unknown label.

Parameters:

• ratings (Any)

• labels (Any)

Return type:

float

Examples

Try it in your browser!

>>> round(fleiss_kappa([[1, 1, 2], [2, 2, 2], [1, 2, 3], [3, 3, 3]]), 4)
0.3617

Go BackOpen In Tab