Might be some interesting opportunities at Kaggle.com for members of the Friam clan.
-tom johnson
---------- Forwarded message ----------
From:
Dan Nguyen <[hidden email]>
Date: Wed, May 9, 2012 at 4:50 AM
Subject: [NICAR-L] Kaggle.com - machine learning contests, datasets, and interviews
To:
[hidden email]I don't know if this has been talked about here recently but I only now found out about Kaggle.com, which is apparently a repository of machine learning and statistical analysis contests
Besides the actual challenges (which include things like thhe Netflix preference predictor), there's a great accompanying blog that asks the winners about their thought process. Their resumes are intimidating but their methods and reasoning are pretty accessible.
Here's one of the better interviews I chanced on last night:
http://blog.kaggle.com/2011/11/23/picture-perfect-bo-yang-on-winning-the-photo-quality-prediction-competition/
The contest involved a set of photos from a photosharing site and designing an algorithm that would predict photo quality based on the metadata alone, including geolocation and tags. The winner talks about how he went as far as to pull in World Bank data, on the reasoning that photos taken in particular population/economic settings may be more appealing to viewers (his hypothesis didn't find much correlation)
The Kaggle main site has some cool datasets to practice with:
http://www.kaggle.com/competitions
Some of them probably contain newsworthy insights on their own, especially if cross referenced with other data sources
====================================================================
To unsubscribe from NICAR-L, please send "unsubscribe NICAR-L" in the
body of an e-mail message to "
[hidden email]". Please
e-mail
[hidden email] if you need help or have questions.
====================================================================
--
==========================================
J. T. Johnson
Institute for Analytic Journalism -- Santa Fe, NM USA
505.577.6482(c) 505.473.9646(h)
Twitter: jtjohnson
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at
http://www.friam.org
Locked
