jshamsho/LFBayes: Bayesian analysis of longitudinal and multidimensional functional data

Longitudinal functional data is functional data repeated over longitudinal time. This package implements the methodology proposed in "Bayesian analysis of longitudinal and multimdensional functional data" by Shamshoian et al. (2020). The use of a weak-separability allows for a computationally scalable algorithm and extracting interpretable marginal covariances and patterns of covariation. The methodology also yields subject-specific latent trajectories and covariate-adjusted mean functions. Simultaneous credible bands are easily available for all quantities of interest.The sampling algorithm is compiled with Rcpp for speed. Regularization is induced by the prior via the Gamma Multiplicative Process. The sampling algorithm is compiled with Cpp for speed gains. Please see the example folder for a quick guide on how to use this package.

install.packages("remotes")
remotes::install_github("jshamsho/LFBayes")