-
The Gap Bootstrap
Fri, Apr 30, 2010 @ 03:00 PM - 04:00 PM
Sonny Astani Department of Civil and Environmental Engineering
Conferences, Lectures, & Seminars
Speakers: Dr. Clifford H. Spiegelman (TAMU), Soumendra Lahiri (TAMU), Justice Appiah (UNL), and Laurence Rilett (UNL)Abstract:In many areas of application, multivariate data are collected routinely over long time periods. Examples include hydrocarbon pollution monitoring, and automated highway volume traffic monitoring. The dominant part or the dependence for these types of data is short term. The gap bootstrap uses a divide, estimate, assess and combine strategy to provide asymptotically optimal or near optimal estimators. In spirit, it is similar in approach to kernel regression estimation, except that the joined pieces are not contiguous in time. We will show that for smooth enough estimators, and some useful dependence models that the resulting estimators are asymptotically efficient and have uncertainties that are accurately assessed using a case bootstrapping approach. Examples will use Origin-Destination (OD) modeling in transportation.
Location: Kaprielian Hall (KAP) - 209
Audiences: Everyone Is Invited
Contact: Evangeline Reyes
