|
|||||
|
|
A functional marked point process model for lupus data |
|
| Authors: | Carlotta Ching Ting Fok James O Ramsay Michal Abrahamowicz Paul Fortin |
| Institution: | 1. Department of Psychology and Center for Alaska Native Health Research, Institute of Arctic Biology, University of Alaska Fairbanks, Fairbanks, AK 99775, USA;2. Professor Emeritus McGill University, 2748 Howe St., Ottawa, Ontario, Canada K2B 6W9;3. James McGill Professor, Research Institute of the McGill University Health Centre, 2155 Guy St., Montreal, Quebec, Canada H3H 2R9;4. Distinguished Senior Investigator of the Arthritis Society at Toronto Western Research Institute, 399 Bathurst St., Toronto, Ontario, Canada |
| Abstract: | Modelling for marked point processes is an important problem, but has received remarkably little attention in the statistical literature. The authors developed a marked point process model that incorporates the use of functional data analysis in a joint estimation of the frequency function of the point process and the intensity of the mark, with application to data from 22 lupus patients consisting of times of flares in symptom severity combined with a quantitative assessment of the severity. The data indicate that a rapid decrease in drug dose is significantly associated with a decrease in flare frequency. Experiments with simulated data designed to model the actual data further support this conclusion. The Canadian Journal of Statistics 40: 517–529; 2012 © 2012 Statistical Society of Canada |
| Keywords: | Frequency function functional data analysis inhomogeneous Poisson process lupus flares marked point process mark variable point process SLEDAI score MSC 2010: Primary 60G55 secondary 62P10 |