Prev
New Competition: R Package Recommendation Engine
What properties determine whether a given R package will be installed by a given user? The people behind Dataists , a new blog for data hackers, want your help to find out. They're hosting a competition with over 120,000 rows of installation information on 52 users and 1,865 R packages. It's their hope that the winning algorithm will be used to create an R package recommendation engine that will help programmers make more informed decisions about which packages to invest time into learning. As well as the recognition associated with building the R recommendations engine, a selection of UseR! books are up for grabs.
Competition Results: Results of the 2010 INFORMS Data Mining Contest
Congratulations to Cole Harris for winning the 2010 INFORMS Data Mining Contest! Cole is a physicist-cum-bioinformatician (and co-founder of Exagen) who spends his time mining medical data to identify genetic features that are diagnostic of disease and predictive of drug response. He will be honored at a session of the INFORMS Annual Meeting in Austin-Texas (November 7-10). Special mentions also to Christopher Hefele (Systems Engineer at AT&T) and Nan Zhou (PhD student at the University of Pittsburgh) who finished 2nd and 3rd out of 147 teams. You can learn more about the winners and their methods on the Kaggle blog.
New Competition: Hearst Challenge
On October 14th, Kaggle will launch a competition for Hearst Magazines on a standalone site. Using a variety of magazine and newsstand sales data, participants are required to predict the optimal number and type of magazines that should be placed on different newsstands. The winner of this competition will receive $25,000.