A Question About Kernel Trick...

Main Page | Report this Page

Computers Forum Index » Computer Artificial Intelligence - Neural Nets » A Question About Kernel Trick...

Page 1 of 1

A Question About Kernel Trick...

Author

Message

DCR...

Posted: Sun Oct 18, 2009 9:50 pm

Guest

If i have a feature expansion:

{0, 1}^n -> {0, 1}^2n. for example. If n = 3, then

Theta(x1,x2,x3) = (1, x1, x2, x3, x1x2, x1x3, x2x3, x1x2x3): Every
possible
"logical OR" from these 3 variables (including 1).

In this case, show K(x,y) for Theta can be computed in polynomial
time: poly
(n).

K(x,y) = Theta(x) dot product Theta(y).

In this case, what are Theta(x) and Theta(y), and more specifically,
how do i solve this problem?

Page 1 of 1

All times are GMT The time now is Mon Dec 07, 2009 9:18 am Computers Forum Index