Department of Statistics Seminar
North Carolina State University

presents

Dr. Michael Sherman

Texas A&M University

"A Statistical Analysis of Tidal Data"

ABSTRACT

In analyzing tidal data we are presented with a regression problem with temporally correlated errors. Ignoring the correlation will be shown to lead to completely incorrect inferences. Loosely, there are two main approaches to account for the correlation: model-based and model-free. In the former the dependence structure is modelled in terms of a few unknown parameters while in the latter the observed series is divided into blocks, and these blocks are used to capture the dependence in the original tidal series. We will compare resulting inferences from model-based and model-free methods in data from Port Mansfield, TX. The model-free approach is quite general applying to other regression problems and to spatial data.

Friday, March 3, 2000

3:35 - 4:35 pm

206 Cox Hall

Refreshments will be served on the second floor of Dabney Hall (left of Room 222) at 3:00 pm.