****************************************************************

************** ICDM 2013: Call for Contest Proposals ***********

****************************************************************

 

We invite proposals for organizers of the ICDM 2013 Data Mining Contest. The Data Mining Contest is an integral part of the ICDM conference and provides an opportunity for teams of scientists and domain experts to compete in order to develop data mining techniques for real-world applications.

Proposals should contain the following information.

 

1.       Contest title and abstract

2.       Contact information for the organizer(s)

3.       General description of the problem

4.       Description of the dataset to be used

5.       Description of the specific competition tasks

6.       Evaluation metrics or tool to be used for evaluation

7.       Plans for contest web site (e.g., kaggle.com, https://c3.ndc.nasa.gov/dashlink, other)

8.       Prize structure

9.       Short biography of organizer(s)

10.   Description of compensation (if any) that could be awarded to winners (organizers are encouraged to solicit awards and prizes from sponsoring entities)

 

General Guidelines

 

·         The dataset must be interesting, available, and large

·         The task(s) must be interesting and accessible to practitioners without domain knowledge

·         The evaluation measures must be well-defined

·         The top contestants must be required to present a paper on their solution at ICDM'13.

·         The contest timeline must adhere to the dates below

 

All submissions should sent by e-mail to [log in to unmask].

 

Important Dates

 

Jun 7: Deadline for submitting proposals for contests

Jun 10: Results announced

Jul 10: Contest begins

Oct 7:  Contest ends

Oct 14: 2-3 page report from each team

Oct 20: camera-ready reports from selected teams.

 

ICDM Contest Co-Chairs

Nitesh Chawla (Notre Dame)

Suresh Venkatasubramanian (University of Utah)
You may leave the list at any time by sending the command

SIGNOFF allstat

to [log in to unmask], leaving the subject line blank.