LFD Book Forum  

Go Back   LFD Book Forum > Course Discussions > Online LFD course > Homework 8

Reply
 
Thread Tools Display Modes
  #1  
Old 05-27-2012, 12:04 PM
markland markland is offline
Member
 
Join Date: Apr 2012
Posts: 30
Default libsvm convergence

Pro tip: I noticed libsvm was giving warnings for some of the runs, saying it hit max iterations. I changed max iterations from 10M to 1B in svm.cpp and recompiled. That prevented the warnings, and the new results affected my homework answers.

Don't know if anyone else has run into this. I'm using svm-train from the command line on Mac OS 10.6.8.
Reply With Quote
  #2  
Old 05-27-2012, 01:08 PM
mikesakiandcp mikesakiandcp is offline
Member
 
Join Date: Apr 2012
Posts: 31
Default Re: libsvm convergence

Nice catch! When I was plotting my values, I realized my answers weren't making sense. Low-and-behold, this was the problem! Thanks!
Reply With Quote
  #3  
Old 05-27-2012, 01:11 PM
rohanag rohanag is offline
Invited Guest
 
Join Date: Apr 2012
Posts: 94
Default Re: libsvm convergence

thanks, this is why I don't submit homework till tuesday
Reply With Quote
  #4  
Old 05-27-2012, 01:39 PM
mic00 mic00 is offline
Invited Guest
 
Join Date: Apr 2012
Posts: 49
Default Re: libsvm convergence

Strange... this changes a few results, but does not change my answers. (I am not bumping up against the limit for the cross-validation questions). Using libsvm through Matlab, OSX 10.7.

EDIT: I take that back. I *occasionally* hit the 10M limit in cross-validation -- it depends on the particular permutation of the data.
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 02:30 PM.


Powered by vBulletin® Version 3.8.3
Copyright ©2000 - 2020, 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.