What do Netflix and the early 18th century British government have in
common? Here’s a hint: it’s not Georgian architecture.
Netflix today announced the Netflix Prize, an award of $1 million
dollars to the first person to improve Netflix’s movie recommendation
system’s accuracy by at least 10 percent.
And according to a statement, the decision to offer the prize was inspired by the
Longitude Prize, created in 1714 by the British government to solve
the problem of how to determine a ship’s longitude.
Netflix will make 100 million anonymous movie ratings ranging from
one to five stars available to contestants.
Currently, when renters watch recommended movies from Netflix, they
rate them on average of a half star higher than the other movies they
rent, Netflix spokesman Steve Swasey told internetnews.com.
Swasey said the Netflix recommendation system uses proprietary
algorithms to process the more than 2 million new ratings a day in order to pair
users with patterns found in the over 1.5 billion ratings already noted.
But if a contestant can bump that half star up another 10 percent,
Netflix will license the technology and publish a
detailed description of the winning approach.
If no one can meet that threshold, the company will annually award a
$50,000 progress prize to whomever makes the most “significant
advancement” toward that goal until someone does meet it.
In order to avoid the scandal that sometimes accompanies the release of
user data, Netflix has removed all personal information from the 100
million ratings, scrubbing them clean of everything but their dates.