DM: Re: your mail

Hi there !

some papers on how to deal with this problem
with Support Vector Machines can be found in:
http://www.support-vector.net/chapter_8.html

where imbalanced datasets are mentioned. In the bioinformatics application
we had quite good results with our method, so you might want to check.

hope this is of help
Nello Cristianini




On Thu, 25 May 2000, Alexey Tsymbal wrote:

 > <200005251105.HAA27209@server10.quality.org>
 > Subject: DM: Re: Classification problem
 > Date: Thu, 25 May 2000 16:47:43 +0300
 > Sender: owner-datamine-l@nautilus-sys.com
 > Precedence: bulk
 > Reply-To: datamine-l@nautilus-sys.com
 >
 >
 > Hello,
 >
 > I had some experience in this kind of problems.
 > It is common especially in medicine, when you often have a few "positive"
 > examples of some
 > disorder.
 > One approach to cope with it is to apply local methods.
 > One good in my opinion example was considered in:
 >
 > Cardie, C., & Howe, N. Improving minority class prediction using
 > case-specific feature weights. Proc. 14th Int. Conf. on Machine Learning,
 > Morgan Kaufmann, pp. 57-65, 1997.
 >
 > Sincerely,
 > Alexey.
 >
 > ___________________________________________________________________
 > Alexey Tsymbal, Researcher                 Department of CS & IS
 > E-mail: alexey@cs.jyu.fi                  University of Jyvaskyla
 > Phone:  +358 14 260 2547                     P.O.Box 35
 > Fax:    +358 14 260 3011                   SF-40351, Jyvaskyla
 > Office: Mattilanniemi D236                      FINLAND
 > WWW: cs.jyu.fi\~alexey
 > ___________________________________________________________________
 >
 >
 > ----- Original Message -----
 > From: "Yannis Kopanas" <ikopanas@ee.upatras.gr>
 > To: <datamine-l@nautilus-sys.com>
 > Sent: Thursday, May 25, 2000 7:17 AM
 > Subject: DM: Classification problem
 >
 >
 >  >
 >  > My problem has to do with the data set. I have two classes (the good guys
 >  > and the bad guys) unfortunatelly the bad guys are only 20 when the good
 > guys
 >  > are 99980. Anybody who knows how to deal with it?
 >  > Thanks in advance.
 >  >      Yannis
 >
 >
 >
 >

• Prev by Date: AW: AW: DM: RE: Data Forms for Mining (Limit on variables)
• Next by Date: DM: RE: Data-mining companies based in London
• Prev by thread: DM: Re: your mail
• Next by thread: DM: Re: Mobile Churn
• Index(es):
  • Date
  • Thread