LFD Book Forum  

Go Back   LFD Book Forum > Book Feedback - Learning From Data > Chapter 2 - Training versus Testing

Reply
 
Thread Tools Display Modes
  #1  
Old 03-17-2013, 07:00 PM
junjy junjy is offline
Junior Member
 
Join Date: Mar 2013
Posts: 2
Question Description on page 55

First, a general comment: Prof. Abu-Mostafa made things really really clear, my million thanks!

Here I have a small confusion: On p.55, line 6, it says "(What the growth function ...), so we can get a factor similar to the '100' in the above example".

The analogy makes the general idea 100 times more comprehensible than plunging into the proof directly. However, here I minded a gap. Can anybody help if this is my misunderstanding or I am right in this point ?

- the '100' is a "good" guy in the above example, which "condenses" (so shrink) the coloured area that times much.
- the growth function, on the other hand, is a bad guy, which gives that much ways for hypotheses behaving differently on the canvas, and "smears" the colours

So I think they are more inversely comparable, e.g. if the example is given as follows:

However many hypotheses in \mathcal{H}, then can only behave in m ways. Therefore, each point on the canvas that is coloured will be coloured M/m times.
Reply With Quote
  #2  
Old 03-17-2013, 11:24 PM
yaser's Avatar
yaser yaser is offline
Caltech
 
Join Date: Aug 2009
Location: Pasadena, California, USA
Posts: 1,477
Default Re: Description on page 55

Quote:
Originally Posted by junjy View Post
First, a general comment: Prof. Abu-Mostafa made things really really clear, my million thanks!

Here I have a small confusion: On p.55, line 6, it says "(What the growth function ...), so we can get a factor similar to the '100' in the above example".

The analogy makes the general idea 100 times more comprehensible than plunging into the proof directly. However, here I minded a gap. Can anybody help if this is my misunderstanding or I am right in this point ?

- the '100' is a "good" guy in the above example, which "condenses" (so shrink) the coloured area that times much.
- the growth function, on the other hand, is a bad guy, which gives that much ways for hypotheses behaving differently on the canvas, and "smears" the colours

So I think they are more inversely comparable, e.g. if the example is given as follows:

However many hypotheses in \mathcal{H}, then can only behave in m ways. Therefore, each point on the canvas that is coloured will be coloured M/m times.
You are correct in the characterization of good and bad guys. The redundancy accounting through factors like 100 is just a technical way to get a handle on the growth function, so indeed they work in opposite directions.
__________________
Where everyone thinks alike, no one thinks very much
Reply With Quote
  #3  
Old 03-19-2013, 01:51 AM
junjy junjy is offline
Junior Member
 
Join Date: Mar 2013
Posts: 2
Default Re: Description on page 55

Thank you for the clearance!
Reply With Quote
Reply

Tags
e_out, growth function

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


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