I have a doubt in inference chapter of deep learning course. Can you please explain how changing the X_binarised_train label from (0,1) to (1,0) increased the accuracy. I am not able to get the proper intutuion of it.
Is it related to the benign and malignant cancer dataset
If yes, that was mainly because there was incorrect assumption made about the dataset about whether (0, 1) represent (malignant, benign) or (benign, malignant).
Once it was identified (bad results gave some idea about something wrong) and corrected(by swapping 0, 1 to 1, 0), results improved.
Alright! makes sense. thank you.