LFD Book Forum

LFD Book Forum (http://book.caltech.edu/bookforum/index.php)
-   Chapter 2 - Training versus Testing (http://book.caltech.edu/bookforum/forumdisplay.php?f=109)
-   -   Exercise 2.13 (http://book.caltech.edu/bookforum/showthread.php?t=4531)

mahaitao 10-23-2014 07:43 PM

Exercise 2.13
 
In Exercise 2.13 (a), Prove dvc(H)<=log_2M. How to think this problem? M is the number of hypotheses, what is the relationship between dvc and M?

(b) What does dvc(\cap H_k) and \cap H_k the intersection of hypotheses mean? How can we intersect hypotheses?

htlin 10-24-2014 10:54 PM

Re: Exercise 2.13
 
For (a), maybe it is worth thinking about the dichotomies that can be generated by M hypotheses?

For (b), the intersection and union of hypothesis "sets" are simply "set" intersection and union.

http://en.wikipedia.org/wiki/Union_%28set_theory%29

Hope this helps.

mahaitao 10-26-2014 04:50 PM

Re: Exercise 2.13
 
Professor Lin,
I think that I did not represent my question clear.

In perceptron example, we have only one H that is a set of infinite lines in the plane. My question is if we consider that H is a union of some subsets, what are they? They are subsets of these infinite lines? How to distinguish them?

And how about intersection of these subsets?

htlin 10-29-2014 01:56 AM

Re: Exercise 2.13
 
For instance, the union of "positive rays" and "negative rays" is "positive or negative rays" which is simply 1-D perceptron. Similarly, you can have perceptrons with w_0 \ge 0, and perceptrons with w_0 \le 0. Their union is all perceptrons; their intersection is perceptrons with w_0 = 0---that is, perceptrons that pass the origin. Hope this helps.

mahaitao 10-30-2014 04:28 PM

Re: Exercise 2.13
 
I got it. Thanks for your patience, professor Lin.


All times are GMT -7. The time now is 03:22 AM.

Powered by vBulletin® Version 3.8.3
Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.
The contents of this forum are to be used ONLY by readers of the Learning From Data book by Yaser S. Abu-Mostafa, Malik Magdon-Ismail, and Hsuan-Tien Lin, and participants in the Learning From Data MOOC by Yaser S. Abu-Mostafa. No part of these contents is to be communicated or made accessible to ANY other person or entity.