Challenges / SIAM SDM'11 Contest: Prediction of Biological Properties of Molecules from Chemical Structure

Status Closed
Type Scientific
Start 2010-11-16 19:00:00 CET
End 2011-01-24 23:00:00 CET
Prize 1,000$

Registration is required.


The final results are now posted (2011-01-25 21:48:13)

Please click "Summary" for the preliminary overview of the final results and "Leaderboard" for ranking.

The QSAR Challenge is now closed (2011-01-24 23:31:19)

Please stay tuned for final results - these will be posted as soon as possible. Thank you for participating!

Last days of the QSAR Challenge (2011-01-22 01:17:47)

Dear Contestants, The SIAM SDM'11 contest is approaching its deadline. This is a friendly reminder that last possible time of submitting a solution is January 24th, 2011, at 23:00:00 CET (Central European Time). Please translate it to your time zone. If in doubt, observe the countdown clock in the Status panel next to "Left" - it's 2 days 21 hours and 48 minutes at the time of writing this reminder. Let us take this opportunity to also remind of other contest rules: 1) Only the Final Test predictions will be taken into account in deciding the contest winner. 2) Only the *last* submission will be considered. 3) In the case of ties, earlier submissions will have higher preference. Wishing you all the good luck, Organizers.

The QSAR Challenge enters new phase (2011-01-10 18:46:13)

Dear Participants, The next phase ("Model Building") of the SIAM SDM'11 contest starts today. The "public" folder contains newly uploaded file "preliminaryTestData.csv" where the 163 Y values for the Preliminary Test Set are revealed. Now, after 2 months of trials you should have an optimal algorithm with best predictivity. Please keep in mind that a method with best *predictivity* does not necessarily mean tailoring/fine tuning your model to crank up its performance on a single (i.e., Preliminary) test set. This issue has been already explained in contest description. Ideally, by now you would have developed new methods of *assessing* predictivity, where you could *estimate* the Balanced Youden Index for Preliminary Test prior to submitting these results - and after submission the number reasonably agreed with what you saw on the Leaderboard. But that would be an ideal situation, the method and manner of model training was and is left entirely up to you. As described in Task section, please combine Preliminary Test Set with Training Data into a single training pool of 837 compounds and develop your final model that will be evaluated on the Final Test predictions. Wishing you all the good luck, Organizers

Happy New Year! (2011-01-06 18:19:46)

The QSAR Challenge is now open. Good luck! (2010-11-16 23:33:35)

Copyright © 2008-2013 by TunedIT
Design by luksite