LFD Book Forum

LFD Book Forum (http://book.caltech.edu/bookforum/index.php)
-   Chapter 4 - Overfitting (http://book.caltech.edu/bookforum/forumdisplay.php?f=111)
-   -   Summary overfit table on page 124 (http://book.caltech.edu/bookforum/showthread.php?t=4789)

ghbcode 10-10-2017 09:57 AM

Summary overfit table on page 124
 
The summary table on the middle of page 124 refers to Figure 4.3.

1. If you increase the number of data points, keeping your H fixed, then you will decrease your overfitting. Got it.

2. If your noise increases, and you keep your H fixed, regardless of the number of data points then overfitting would decrease, no? Why would overfit increase? Isn't overfit usually a function of and overly complex hypothesis?

3. If your target function complexity increases, keeping all else constant, wouldn't that lead to less overfit?

In my opinion this table is not very helpful and neither is figure 4.3. Is there a better way to think about this? In my opinion there should be a third column to better illustrate what is fixed, what is changing and what the end result is


All times are GMT -7. The time now is 06:22 PM.

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.