The netflix prize
November 20th, 2007
Eliezer Hernandez (Senior, CS major), Michael Showatter (Sophomore, Math Ed major) and I formed a team called “EMU” to compete for the netflix prize.
Netflix offers one million dollar for an algorithm which is able to improve the performance of their current movie recommendation algorithm by 10%. They use the RMSE score to measure the performance of an algorithm. Any algorithm that produces a RMSE score of lower than or equal to 0.8563 will win.
We have tried out some simple approaches. We have lowered our RMSE score from the original 1.7686 to 1.0536 (Good job! Eliezer and Michael.). The team of researchers from AT&T labs has the current best score 0.8705. For now we are trying to understand and implement their ideas. We should be able to spend more time on it during the winter break.
I have to say the chance for us to win the prize is slim. The important thing is not to win the money but the valuable experiences one gains along the way. Plus, it is a nice feeling to compete with teams all around the world, probably from every top university. And if we do win, we promise to donate at least half of the money to help build the new science center :-).
In the future I would like to see more math/cs students exhibiting such enthusiasm and energy.
