|
Activity Number:
|
353
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 14, 2002 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics & the Environment*
|
| Abstract - #301527 |
|
Title:
|
Bayesian Methods for Circular Regression Using Wrapped Distributions
|
|
Author(s):
|
Palanikumar Ravindran*+ and Sujit Ghosh
|
|
Affiliation(s):
|
North Carolina State University and North Carolina State University
|
|
Address:
|
1707 Crest Road Apt 5, Raleigh, North Carolina, 27606, USA
|
|
Keywords:
|
Regression ; Circular data ; Wind direction ; bayesian ; Wrapped normal ; Wrapped double exponential
|
|
Abstract:
|
Circular data arise in a number of different areas in environmental sciences like geological and meteorological sciences. Some common examples are wind directions, ocean current directions, and direction of movement of hurricanes. The standard regression methods can not be used to model circular data because of the circular geometry of the sample space. One of the common methods to analyze such data is the wrapping approach. In this approach, we assume that the probability distribution for circular data is obtained by wrapping a probability distribution on the real line onto a circle. However, the likelihood-based inference for such distributions can be very complicated and computationally intensive. We discuss how Markov Chain Monte Carlo (MCMC) methods using a data augmentation step are flexible and computationally efficient to fit a wide class of wrapped distributions. We apply our method to a wind direction data set and compare our results to parameter estimates available in the literature. Simulations are presented to validate the proposed method.
|
