Hi,

In the last part of the programming assignment, we are asked to compute "How many of the support vectors lie on the margin hyperplane?".

1. When you refer to "alpha_n = C", I guess you didn't mean the regularization C because those seem unrelated. So what exactly did you mean?

2. Assuming I have the size of the margin (let's denote that as M on each side of the separating plane), should I return all points that are within distance < M from the hyperplane (i.e. those that have a non-zero slack variable), or all the points with distance = M (i.e. no need for slack variable)? Or maybe sum of both?

Thanks in advance!

~ Barak