very nice video ! You helped me very much, thank you. " ξ " is not spelled zeta but like this translate.google.gr/#el/en/%CE%BE from a Greek fan of your videos :D
@7:58 How do we compute beta0? In non kernel version we get beta0 by putting value of beta(transpose) on the equation of line y(beta(trans)* xi + beta) = 1. Any ideas?
Beta0 is a singular value . So we can only calculate it in the non kernel version . It will remain the same even when the data are projected to higher dimensions. Since beta0 is a single value we cannot use the kernel to change its dimensions
Best Lecture ever. I am a fan of your lectures and sometimes I watch them several times.
One of the best videos i have come across on SOFT margin.
Very nicely explained. Subtle mathematics and sufficient explanation about it.
Great lecture
very nice video ! You helped me very much, thank you. " ξ " is not spelled zeta but like this translate.google.gr/#el/en/%CE%BE from a Greek fan of your videos :D
@7:58 How do we compute beta0? In non kernel version we get beta0 by putting value of beta(transpose) on the equation of line y(beta(trans)* xi + beta) = 1. Any ideas?
Beta0 is a singular value . So we can only calculate it in the non kernel version . It will remain the same even when the data are projected to higher dimensions. Since beta0 is a single value we cannot use the kernel to change its dimensions