1
$\begingroup$

I'm evaluating the performance of a trained neural network, and I'm getting the following stats:

AUC-ROC: 0.91

accuracy: 0.85

How should I interpret this result?

I'll add that the accuracy is a result of one-hotting the output and comparing it against the label, where as the AUC-ROC is calculated using the probabilistic (ie. range 0->1) output of a label. I'm not sure if it is correct to compare the metrics when doing this.

Improve this question
$\endgroup$
0

1 Answer 1

Reset to default
1
$\begingroup$

The overall accuracy is measured at a specific point, and reflects your model's ability to place data into a proper category based on a specific threshold.

AUC more so reflects your model's ability to show a constant difference in data with different labels (i.e class A is generally higher/lower than class B), and doesn't explicitly depend upon the model's ability to correctly assign things to classes.

Improve this answer
$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.