|
Activity Number:
|
363
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #300563 |
|
Title:
|
Joint Spatial Modeling of Recurrent Tree Infection and Growth with Processes Under Intermittent Observation
|
|
Author(s):
|
Farouk S. Nathoo*+
|
|
Companies:
|
University of Victoria
|
|
Address:
|
PO Box 3045 STN CSC, Mathematics and Statistics, Victoria, BC, V8W3P4, Canada
|
|
Keywords:
|
recurrent events ; panel data ; growth curves ; multivariate spatial process ; kernel convolution ; joint modelling
|
|
Abstract:
|
In this article we present new methods for longitudinal studies in forestry where trees are subject to recurrent infection and the hazard of infection depends on tree growth over time. Here, tree height is an intermittently observed, time-dependent and spatially-varying covariate that is subject to measurement error. We develop a joint model linking a counting process for recurrent infections with a nonlinear spatial growth curve representing the underlying height trajectories. To accommodate spatial variability in growth parameters corresponding to trees at different locations, a flexible and robust non-Gaussian multivariate spatial process model is developed. We adopt an approach based on kernel convolution with the kernel modeled in a semiparametric fashion. In particular, a penalized smoothing approach based on a piecewise constant kernel and Markov random field priors is employed.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2008 program |
