NCRN Virtual Seminar February 5, 2014

The NSF-Census Research Network is eight nodes comprised of researchers conducting innovative, high-disciplinary investigations of theory, methodology and computational tools of interest.

NCRN conducts a virtual seminar at one of the nodes, broadcast to all the others and to federal statistical agencies, every month on the first Wednesday from 3:00-4:30. Information on this month's seminar (originating from Cornell) is below.

NCRN Virtual Seminar: Boosting Models for Edit, Imputation and Prediction of Multiple Response Outcomes

Wednesday, February 05, 2014, 3:00PM - 4:30PM

Speakers: Ping Li (Cornell and Rutgers) and John Abowd (Cornell)


In this paper, we propose a statistical framework that generalizes the classical logit model to predict multiple responses (i.e., multi-label classification). We develop an effective implementation based on boosting and trees. For the NCRN seminar we present an application to editing and imputation in the multiple response race and ethnicity coding on the American Community Survey.