The R Recommendation Engine contest is now live on Kaggle. Please head over there and start submitting your predictions for the test data set. Once you do, you can check the leaderboard to see how your algorithm compares with other people’s work. We know that there’s still plenty of progress that can be made, because we have other models that are much better than the benchmark code we released to the public.
In the future, I’ll be posting some hints on this blog about ways to improve your models for the contest using less well-known methods in machine learning. Because you can use CRAN itself as a data source, this contest offers a lot of opportunities to exploit the state-of-the-art in machine learning based on text and network analysis.
If you any questions or comments about the contest, please use the contest forum on Kaggle so that others will benefit from the discussion.