Support Vector Machines - The Auton Lab
Support Vector Machines - The Auton Lab
Support Vector Machines - The Auton Lab
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
Learning Maximum Margin with Noise<br />
wx+b=1<br />
wx+b=0<br />
wx+b=-1<br />
ε 2<br />
ε 11<br />
ε 7<br />
M =<br />
2<br />
w.w<br />
Given guess of w , b we can<br />
• Compute sum of distances<br />
of points to their correct<br />
zones<br />
• Compute the margin width<br />
Assume R datapoints, each<br />
(x k ,y k ) where y k = +/- 1<br />
What should our quadratic<br />
optimization criterion be?<br />
Minimize<br />
1<br />
2<br />
w.<br />
w + C<br />
R<br />
∑ ε k<br />
k = 1<br />
How many constraints will we<br />
have? R<br />
What should they be?<br />
w . x k + b >= 1-ε k if y k = 1<br />
w . x k + b = 1-ε k if y k = 1<br />
w . x k + b