[Rd] e1071 ver 1.5-27 and older - SVM bug report

From: David Meyer <meyerd_at_technikum-wien.at>
Date: Sat, 06 Aug 2011 12:38:11 +0200


Dear Pawel:

yes, this is a bug. Fixed in the next release.

Thanks

David


Dear All:

I found a problem with the SVM internal cross-validation (CV) accuracy estimation in the e1071 package.

File: Rsvm.c
Line: 120

Today, it is:

int j = rand()%(prob->l-i);

Should be:

int j = i + rand()%(prob->l-i);

The erroneous code doesn't shuffle objects. Instead, it "randomly" moves objects from beginning to the end.



R-devel_at_r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel Received on Sat 06 Aug 2011 - 12:45:13 GMT

This quarter's messages: by month, or sorted: [ by date ] [ by thread ] [ by subject ] [ by author ]

All messages

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Sat 06 Aug 2011 - 21:00:14 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-devel. Please read the posting guide before posting to the list.

list of date sections of archive