Longitudinal Models in PyMC
Longitudinal Analysis of Growth Trajectories
In this project I outline the strategy of using multi-level or hierarchical Bayesian models to evaluate the growth trajectories of individuals, and estimate the between individual effects. The main theme of the work is to try and disambiguate some of the complexities of mixed level (hierarchical modelling) when applied to longitudinal data.
The project culminated in a publication to the official PyMC documentation that can be found online here and downloaded here
The notebook demonstates these techniques applied to data of youth alcohol consuption. and behaviourial data. In particular we show how Bayesian hierarchical methods can be used to help characterise the manner in which different factors such can influence an individual’s trajectory and why it’s important to understand the sources of variation in such growth trajectories.
