Support = 'None'

Question

Using the code below I get values for precision, recall, and F scores but I get None for support

import numpy as np
from sklearn.metrics import precision_recall_fscore_support
ytrue = np.array(['1', '1', '1', '1', '1','1','1','1','0'])
ypred = np.array(['0', '0', '0', '1', '1','1','1','1','0'])
precision_recall_fscore_support(ytrue, ypred, average='weighted')

output:

(0.91666666666666663, 0.66666666666666663, 0.72820512820512828, None)

I checked http://scikit-learn.org/stable/modules/generated/sklearn.metrics.precision_recall_fscore_support.html but I find it a bit unclear as to why it is None

Questions:

1. Why is support equal to None in my output?
2. How do I get a non-None output?

Answer 1

Why is support equal to None in my output?

If a value for average is provided, None is returned for support

How do I get a non-None output?

Don't provide a value for average. If you still want to use weighted and need the support, just do something like

> from sklearn.metrics import confusion_matrix
> np.sum(confusion_matrix(ytrue, ypred), axis=1)
array([1, 8])