This preview shows page 1. Sign up to view the full content.
Unformatted text preview: nce can be expressed in terms of dot
products: Dis 2 (x, y) = ||x y||2 = p (x y )| (x Replacing dot products with kernels: p
Dis K (x, y) = K (x, x) y) = p 2K (x, y) x| x 2x| y Replacing dot products with kernels: Dis K (x, y) = y| y p
K (x, x) 2K (x, y) K (y , y ) If we consider a kernel that satisfies K(x,x) = 1, then nearest
neighbor classification with kernels or distances is equivalent. K (y , y ) As an alternative, consider kernels as measures of similarity,
and rather than looking for the closest points, look for the most
similar points, and use kernels directly. 29 30 8...
View Full Document
This note was uploaded on 02/10/2014 for the course CS 545 taught by Professor Anderson,c during the Fall '08 term at Colorado State.
- Fall '08
- Machine Learning