LFD Book Forum (http://book.caltech.edu/bookforum/index.php)
-   Chapter 2 - Training versus Testing (http://book.caltech.edu/bookforum/forumdisplay.php?f=109)

 Sangrock Lee 09-28-2016 07:42 PM

problem 2.14-(b)

Hi,

From the given formula, I can derive d_vc<l, but I don't know how to make this d_vc <= l. Can anyone give me a clue on it?

Thanks.

 Haoming 10-02-2016 08:06 PM

Re: problem 2.14-(b)

Quote:
 Originally Posted by Sangrock Lee (Post 12440) Hi, From the given formula, I can derive d_vc
I have the same issue as yours.

The way I approach this problem is that, since exp(l) > poly(l), l must be greater than break point. But I have a strong feeling that this is not the right way to do it.

 Wennan 10-04-2016 06:29 PM

Re: problem 2.14-(b)

I also get d_vc < l. Actually d_vc < l implies d_vc <= l. So it works if we get that.

Quote:
 Originally Posted by Sangrock Lee (Post 12440) Hi, From the given formula, I can derive d_vc

 dubwub 10-05-2016 01:11 PM

Re: problem 2.14-(b)

How did you prove that d_VC(H) < l? I've proved that d_VC < l but I'm having difficulty expanding it to the entire union hypothesis set.

EDIT: Oh just kidding, figured it out, part a is helpful

 williamshakespare 11-27-2016 02:56 AM

Re: problem 2.14-(b)

Quote:
 Originally Posted by dubwub (Post 12449) How did you prove that d_VC(H) < l? I've proved that d_VC < l but I'm having difficulty expanding it to the entire union hypothesis set. EDIT: Oh just kidding, figured it out, part a is helpful
how do you induce dvc(H) < l? I still can't make it.
Thanks a lot

 williamshakespare 11-27-2016 03:10 AM

Re: problem 2.14-(b)

Quote:
 Originally Posted by dubwub (Post 12449) How did you prove that d_VC(H) < l? I've proved that d_VC < l but I'm having difficulty expanding it to the entire union hypothesis set. EDIT: Oh just kidding, figured it out, part a is helpful
how do you induce dvc(H) < l? I still can't make it.
Thanks a lot

 All times are GMT -7. The time now is 08:00 PM.