Shouldn’t this be y_hat = 0 inside the curve and y_hat = 1 outside the curve?
Thanks in advance 
It depends on what the ‘y’ values are for each example.
If the ‘y = 1’ values are outside the ellipse, that’s where’d you’d expect the h-hat values also.
The curve just tells you where the 0.5 values are - it separates the examples. It doesn’t by itself tell you which region is which.
1 Like
Got it… Thanks.
Could you also please provide the mathematical proof for the cost function of logistic regression to be convex?
Can be found online (via stackexchange)
1 Like