Experts will convene at the ACM Conference on Recommender Systems, which takes place Oct. 22-25 in New York City, to discuss what insights have been gleaned from the Netflix Prize, a $1 million competition to improve the accuracy of the movie rental company’s in-house recommendation system by 10 percent. Ces Bertino, a member of one of the teams that competed for the Netflix Prize, says the contest had value in forcing participants to apply all algorithms to the same set of uneven real-world data rather than choosing the datasets. “Because people had to use a fixed dataset, they needed to deal not only with the advantages of a particular method, but also the weaknesses of it,” Bertino says. “You could not escape it.” Another contestant, Gavin Potter, says the winning algorithms succeeded through the recognition that combining various strategies generated the best results. University of Minnesota professor John Riedl says the winning models imply that integrating many algorithms with machine-learning methods could conceivably be a solid general strategy for handling large datasets. He believes it is time for people in the field of recommendation systems to focus on neglected system elements that could benefit the industry. The winning algorithms also could find application in areas that include fraud detection, market trading, fighting spam, and computer security, says Nicholas Ampazis at Greece’s University of the Aegean.
For More Information:http://www.cpccci.com
This entry was posted on Friday, October 9th, 2009 at 7:04 pm and is filed under Business Internet News. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.