supress value error when label contains one label #157
Conversation
Codecov Report
@@ Coverage Diff @@
## master #157 +/- ##
==========================================
+ Coverage 77.92% 78.57% +0.65%
==========================================
Files 79 80 +1
Lines 3388 3468 +80
==========================================
+ Hits 2640 2725 +85
+ Misses 748 743 -5 |
| @@ -73,6 +74,10 @@ class ROCAUCEvaluator(Evaluator): | |||
| are considered as negative. | |||
| ignore_labels (int or list or None): labels to be ignored. | |||
| `None` is used to not ignore all labels. | |||
| raise_value_error (bool): If False, `ValueError` caused by | |||
| `roc_auc_score` calculation is suppressed and ignored with warning | |||
| @@ -73,6 +74,10 @@ class ROCAUCEvaluator(Evaluator): | |||
| are considered as negative. | |||
| ignore_labels (int or list or None): labels to be ignored. | |||
| `None` is used to not ignore all labels. | |||
| raise_value_error (bool): If False, `ValueError` caused by | |||
There was a problem hiding this comment.
enclose False with backquotes.
| else: | ||
| # This is usually caused by the following | ||
| # Only one class present in y_true. | ||
| # ROC AUC score is not defined in that case |
There was a problem hiding this comment.
It describe when roc_auc_score is raised. So I think we should move to l.145.
|
updated |
| try: | ||
| roc_auc = metrics.roc_auc_score(t_total, y_total) | ||
| except ValueError as e: | ||
| # `ValueError` is usually caused by the following |
There was a problem hiding this comment.
It seems you list only one situation where ValueError is raised. So, by the following seems strange.
| roc_auc = metrics.roc_auc_score(t_total, y_total) | ||
| except ValueError as e: | ||
| # `ValueError` is usually caused by the following | ||
| # Only one class present in y_true. |
There was a problem hiding this comment.
We should not capitalize here because the previous sentence is not completed.
|
ROCAUCEvaluator will raise ValueError when only one label is contained in the dataset. It has another benefit that we can make test example faster. |
|
Could you review @delta2323 @mottodora ? |
|
LGTM |
No description provided.