LFD Book Forum

LFD Book Forum (http://book.caltech.edu/bookforum/index.php)
-   Chapter 1 - The Learning Problem (http://book.caltech.edu/bookforum/forumdisplay.php?f=108)
-   -   Exercise 1.10d (http://book.caltech.edu/bookforum/showthread.php?t=4802)

saitama 12-06-2017 08:28 AM

Exercise 1.10d
 
I can see why coin c_1 obeys the Hoeffding bound since we restrict our attention to a single coin (eg a single bin). And i can understand why c_min does not hold since we pick an optimal value (ie minimum) over multiple coins (kind of like the best hypothesis over multiple bins).

However, I'm having a hard time understanding why the Hoeffding bound also applies to the c_rand coin. Although we don't pick an optimal value we do consider multiple coins (eg multiple bins). Does considering multiple coins not matter in this case??

don slowik 12-06-2017 11:01 AM

Re: Exercise 1.10d
 
Just one bin is being considered after you have chosen it. Just like choosing bin 1..
Hoeffding should apply to any one of the 100,000 runs, but not to the 'best' or 'worst' run.

saitama 12-08-2017 06:57 AM

Re: Exercise 1.10d
 
I think I understand now. I guess the c_rand coin does only considering one bin at a time when choosing an optimal value. Its not like we search multiple coins. Thanks for the reply


All times are GMT -7. The time now is 04:33 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.