Smooth random change point models

Van Den Hout, Ardo; Muniz-Terrera, Graciela; Matthews, Fiona E.

Abstract

Change point models are used to describe processes over time that show a change in direction. An example of such a process is cognitive ability, where a decline a few years before death is sometimes observed. A broken-stick model consists of two linear parts and a breakpoint where the two lines intersect. Alternatively, models can be formulated that imply a smooth change between the two linear parts. Change point models can be extended by adding random effects to account for variability between subjects. A new smooth change point model is introduced and examples are presented that show how change point models can be estimated using functions in R for mixed-effects models. The Bayesian inference using WinBUGS is also discussed. The methods are illustrated using data from a population-based longitudinal study of ageing, the Cambridge City over 75 Cohort Study. The aim is to identify how many years before death individuals experience a change in the rate of decline of their cognitive ability. © 2010 John Wiley & Sons, Ltd.

Citation

Van Den Hout, A., Muniz-Terrera, G., & Matthews, F. E. (2011). Smooth random change point models. Statistics in Medicine, 30(6), 599-610. https://doi.org/10.1002/sim.4127

• Journal Article Type: Article
• Publication Date: Mar 15, 2011
• Deposit Date: Dec 8, 2023
• Journal: Statistics in Medicine
• Print ISSN: 0277-6715
• Electronic ISSN: 1097-0258
• Publisher: John Wiley and Sons
• Volume: 30
• Issue: 6
• Pages: 599-610
• DOI: https://doi.org/10.1002/sim.4127
• Public URL: https://hull-repository.worktribe.com/output/4454918