Go to  Advanced Search

Stocahastic process based regression modeling of time-to-event data : application to phenological data

Show full item record

Files in this item

Files Size Format Description   View
ubc_2010_fall_cai_song.pdf 875.5Kb Adobe Portable Document Format   View/Open
Title: Stocahastic process based regression modeling of time-to-event data : application to phenological data
Author: Cai, Song
Degree: Master of Science - MSc
Program: Statistics
Copyright Date: 2010
Issue Date: 2010-08-23
Publisher University of British Columbia
Abstract: In agricultural study, the timings of phenological events, such as bud-bursting, blooming and fruiting, are considered to be mainly influenced by climate variables, especially accumulative daily average temperatures. We developed a stochastic process-based regression model to study the complicated relationship between phenological events and climate variables, and to predict the future phenological events. Compared with the traditional Cox model, the newly developed model is more efficient by using all available time-dependent covariate information, and is suitable for making predictions. Compared with parametric proportional hazards model, this model is less restrictive on assumptions, and fitting of this model to data is computationally straightforward. Also, this model may be easily extended to incorporate sequential events as responses. It may also be useful for a broad range of survival data in medical study. This model was applied to the bloom-date data of six high-valued, woody perennial crops in the Okanagan Valley, BC Canada. Simulation results showed that the model provides a sensible way to estimate an important parameter, Tbase, controlling phenological forcing events. Also, our statistical findings support Scientists' previous experimental findings that the temperature influence blooming events via accumulation of growing degree days (GDDs). Furthermore, a cross-validation procedure showed that this model can provide accurate predictions for future blooming events.
Affiliation: Science, Faculty of
URI: http://hdl.handle.net/2429/27645
Scholarly Level: Graduate

This item appears in the following Collection(s)

Show full item record

Attribution-NonCommercial 2.5 Canada Except where otherwise noted, this item's license is described as Attribution-NonCommercial 2.5 Canada

UBC Library
1961 East Mall
Vancouver, B.C.
Canada V6T 1Z1
Tel: 604-822-6375
Fax: 604-822-3893