Features

After a rapid progress in the first 3 weeks, the progress in the Netflix Prize, as measured by the decrease in RMSE, has dramatically slowed down -

see http://www.netflixprize.com/community/viewtopic.php?id=386

For about two weeks (as of Nov 6), the best results seem to be stuck just below 5% improvement over the baseline, still far from the 10% improvement required for the prize.

There are several possible explanations, including
1) Slowdown is just temporary
2) Leaders do not want to post their best results (saving for later)
3) Slowdown is due to pause while competitors get movie information (not included in Netflix data)
4) New algorithms are needed - current algorithms reached their peak
5) There is a limit to prediction accuracy (due to the random component of ratings) and the leaders are getting close to that limit

When do you think 10% improvement in RMSE, required for Netflix Grand Prize, will be achieved?

Please vote on www.kdnuggets.com and choose from

• Nov 06
• Dec 06
• Q1 07
• Q2 07
• Q3-Q4 07
• 2008
• 2009 and later
• Never