AAAI 2000 Workshop

Instructions for Authors


Papers accepted for presentation at the AAAI-2000 workshop on Learning from Imbalanced Data Sets will be included in the Workshop's Working Notes as well as the Workshop's Technical Report (see Section 1. of "Additional Notes" below for a description of these publications). Further formal publication avenues for the papers or extended versions of them will be discussed at the workshop.

Submission of Camera-Ready Copies to Organizers

Towards the end of April, AAAI will be mailing you and your co-authors workshop registration materials and A/V forms. Several documents must be received by the organizers of the workshop on or before Wednesday April 26, 2000 . These documents are:

To be sent by Physical Mail (see physical address below):

  • A final camera-ready version of your paper
  • A completed copy of the A/V form
  • A completed and signed copy of the "permission to distribute" form that can be found here (NB: signing this form does not take away your right to publish your work in another forum such as in a conference proceedings or in a journal).

To be submitted electronically:

  • An electronic version of your paper (see electronic address below)
  • An electronic abstracts of your paper using the form located here (see Section 2. of "Additional Notes" below for further instructions)

Paper Style

In order to conform to AAAI's formatting style, please follow the AAAI author's instructions described here to prepare the camera-ready and e-mail version of your paper.


Physical Version

Nathalie Japkowicz
Faculty of Computer Science
DalTech/Dalhousie University
6050 University Avenue
Halifax, N.S.
Canada, B3H 1W5

Electronic Version

Deadline for Submission of the Final Paper and Accompanying Documents

All physical and electronic documents must be received at the addresses above on or before Wednesday April 26, 2000

AAAI 2000 Workshop

Additional Notes

1. Technical Report:

The workshop papers will be made available at the workshop in the form of Working Notes as well as after the workshop in the form of a AAAI Technical Report .

Working Notes are available only at the workshop and they are not citable. Although issuance of a technical report does not constitute formal publication, it does provide authors with a mechanism for citing their work. Furthermore, Technical Reports are available to anyone from the AAAI Press a short time (usually, several months) after the conference.

2. Electronic Abstracts of Papers

Workshop participants must submit electronic abstracts of their paper to AAAI by April 26, 2000. Participants should use the form located here

The event to choose from the pull-down list is "AAAI Workshop"

It is extremely important that the Workshop title also be given, using the code W10, which stands for the following workshop:

W10: Learning from Imbalanced Data Sets

Participants who are unable to use the web-based form should submit their abstracts to the workshop organizer.

3. Electronic Versions of Papers

Electronic versions of all the papers should be e-mailed to by April 26, 2000 in PDF or PostScript files which should be set up using letter size (8.5 x 11 inches) as the default page size--not A4, and all fonts must be embedded. We recommend that authors use Times, Symbol, and/or Helvetica fonts.

To avoid possible erasure of files on the AAAI site, participants should name their papers with the first initial followed by the last name of the primary author, along with the year of the workshop and the extension .pdf or .ps (depending upon the file type). Thus, a paper by John Doe would be called JDoe00.pdf or

AAAI requests electronic versions of the papers because it is in the process of archiving all AAAI Press publications. At some point, all the publications will be available to AAAI members on the Web (Proceedings papers will be available this year. Posting of all workshop papers and symposia papers may take a bit longer).

Unofficially, the workshop papers will all be made available to anyone through the Workshop's Web Page at this address.

4. CoIL Competition

We would like to point to your attention that the CoIL competition data set this year is a 2-class unbalanced data set -- 5000 in the majority class and 300 in the minority class. See challenge

[This is only for your information: for the purpose of the workshop, it is not required that your system be run on this data set]