From what I understood, it's not that the decision boundary corresponds to the 2nd PCA variable. It's that you reduce your 2D data set to 1D by projecting it onto the first PCA variable. So after, imagine that you have only a line and all the points are projected onto this line (which is the 1rst PCA). So then, on this line, for the first data set, you can easily use a threshold and separate the points, while in the second data set you cannot. Hope that helps!
Exam2019 pb21
Hello,
I am not sure why in pb21 of the 2019 exam do we have that the decisionboundary corresponds to the 2nd PCA variable?
Thanks a lot!
1
From what I understood, it's not that the decision boundary corresponds to the 2nd PCA variable. It's that you reduce your 2D data set to 1D by projecting it onto the first PCA variable. So after, imagine that you have only a line and all the points are projected onto this line (which is the 1rst PCA). So then, on this line, for the first data set, you can easily use a threshold and separate the points, while in the second data set you cannot. Hope that helps!
2
Add comment