Mudd Alumnus Robert Bell Helps Solve Netflix’s $1 million Question

Robert Bell HM ’72 is a member of an international team that recently won the $1 million Netflix prize for creating the best recommendation software to predict trends in Netflix customer movie preferences.Three years ago, Netflix issued a worldwide challenge to engineers, scientists, and statisticians, asking them to improve Netflix’s ability to accurately predict what movies their customers would like by 10 percent. Bell’s team beat thousands of others, submitting their algorithm a mere 20 minutes before the July deadline.“BellKor’s Pragmatic Chaos,” as the seven-person team is called, is made up of employees from AT&T Labs in New Jersey and BigChaos of Austria. Bell works in the Statistics Research Group at AT&T Labs.The team said the algorithm made advances in two important fields of taste prediction. Their “neighborhood modeling” improved methods for finding films that share certain characteristics, while their other approach focused on “latent factors”—that is, movie characteristics that are identified mathematically rather than by human evaluation.After graduating from Harvey Mudd College as a mathematics major, Bell went on to earn his Master’s in Statistics at the University of Chicago and his Ph.D. in Statistics at Stanford University. After earning his Ph.D., he worked for the Rand Corporation doing public policy analysis until joining AT&T Labs’ Statistics Research Department in September 1998. His current research involves data analysis and model building. He also serves on the board of the National Institute of Statistical Sciences and is a fellow of the American Statistical Association.

Facebook Comments

Leave a Reply