Talk:Support vector machine/Archives/2014

This is an archive of past discussions about Support vector machine. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Does the solution lie at a saddle point?

The "Primal form" section says that "we look for a saddle point". Is this correct? What is thefunction that this is a saddle point of? Is it a saddle point of $f(w,b)=\max _{\alpha }{\frac {1}{2}}||w||^{2}+\sum _{i}\alpha _{i}(y_{i}(w_{i}x_{i}-b)-1)$ ? It doesn't seem clear that this is even differentiable. The objective function (the norm squared of (w,b)) is positive semi-definite, so I don't think it has any saddle points.Vinzklorthos (talk) 01:37, 7 February 2014 (UTC)

Contradiction in primal form?

IMHO, the section on the determination of $b$ at the end of the primal form section is unclear and seems contradictory. More precisely, the statement "the b depends on y_i and x_i" does not seem to follow from anything previously said. On the contrary, the previous statement "From this one can derive that the support vectors also satisfy ..." refers to a single value of $w$ and $b$ for which the given equations hold.

Veryltdbeard (talk) 11:11, 27 October 2014 (UTC)