Netflix Cancels Recommendation Engine Contest, Settles Privacy Lawsuit [NewTeeVee]
Over three years ago, Netflix launched a contest with a $1 million dollar prize. The contest winner would be the person or team that wrote an algorithm that improved the company's movie recommendations based on previous movie rentals by 10 percent or more. In order to provide the contestants with data to test their recommendation algorithms, Netflix released a data set of actual customers rental information, with names and identifying information removed.
Ultimately, one team of computer scientists wrote an algorithm that improved Netflix's recommendations by 10.5 percent and was awarded the price money. The improved algorithm provides customers reserving DVDs from the Netflix web site and Netflix streaming customers recommendations they are more likely to enjoy.
But, a FCC investigation has halted a second contest designed to further improve the company's recommendations because of privacy concerns about the release of a second data set of customer rental habits. The second data set was to have included some demographic data of individual Netflix customers.
Netflix has announced that it has terminated the second round of the contest and settled a related lawsuit that had been filed by Netflix users concerned about the privacy of their personal information. For now, Netflix will rely on its own employees to improve its recommendations rather than crowdsourcing algorithm assistance.
Comments