lesson03: binary confusion matrix true/false order

[elcorto]

In addition to deeplearning540/lesson03#1, sklearn also uses y_{true,pred} = [False, True] such that TN = C[0,0] and so forth, as opposed to [True,False] in wikipedia.

https://scikit-learn.org/stable/modules/generated/sklearn.metrics.confusion_matrix.html

[psteinb]

Sounds like we simply should create our own table. :-) That is long overdue.

[PrometheusPi]

Could it be that the labels used in the lecture video for ConfusionMatrixDisplay were inverted?

If tp = cm[1,1] that the lower right entry should be the Adele penguin. But with display_labels=["Adele", "not Adele"] it looks like the upper left (true negative) is the Adele penguin.

Wouldn't display_labels=["not Adele", "Adele"] be the correct labeling?