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 04-28-2012, 07:02 PM
pventurelli@gotoibr.com pventurelli@gotoibr.com is offline
Member
 
Join Date: Apr 2012
Posts: 12
Default Questions on Probability

I have a question on the use of the Expectation Value operation as used in the class and the book.

My (simplistic) understanding of the expectation value of a quantity is basically like taking a weighted average: for random variable X with probability distribution Pr(X) the expectation value would be either a sum (if X is discrete) or an integral (X is continuous) over all X values of X_i * Pr(X=X_i). And then I would multiply everything by a 1/N or a 1/Delta(X).

Is this more or less what is meant in the class by the E[] notation? Maybe I just need to become more comfortable using this operation, however language like that on page 61 of the book where the in sample error is defined as an averaging procedure in (implied?) contrast to the definition of E_out leaves me confused. Isn't expectation a weighted average *by definition*?

Thanks for your help!
Reply With Quote
  #2  
Old 04-28-2012, 07:38 PM
yaser's Avatar
yaser yaser is offline
Caltech
 
Join Date: Aug 2009
Location: Pasadena, California, USA
Posts: 1,475
Default Re: Questions on Probability

Quote:
Originally Posted by pventurelli@gotoibr.com View Post
I have a question on the use of the Expectation Value operation as used in the class and the book.

My (simplistic) understanding of the expectation value of a quantity is basically like taking a weighted average: for random variable X with probability distribution Pr(X) the expectation value would be either a sum (if X is discrete) or an integral (X is continuous) over all X values of X_i * Pr(X=X_i). And then I would multiply everything by a 1/N or a 1/Delta(X).

Is this more or less what is meant in the class by the E[] notation? Maybe I just need to become more comfortable using this operation, however language like that on page 61 of the book where the in sample error is defined as an averaging procedure in (implied?) contrast to the definition of E_out leaves me confused. Isn't expectation a weighted average *by definition*?

Thanks for your help!
The sample average can be technically posed as an expected value (by giving each example a probability 1/N). However, the correct way of thinking about E_{\rm in} is that it is a random variable (the source of randomization being which data set was generated) rather than an expected value like E_{\rm out}. However it is a random variable obtained by averaging a (hopefully) large number of sample points, and as such its value will tend to be close to the expected value E_{\rm out} per the laws of large numbers.
__________________
Where everyone thinks alike, no one thinks very much
Reply With Quote
Reply

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 03:19 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.