spaMM (version 2.2.0)

sparse_precision: Sparse_precision algorithm

Description

A fitting algorithm efficient for random effects with sparse precision matrix (i.e. inverse covariance matrix) is implemented. It is used by default only in two cases: for conditional autoregressive models (with a random effect of the form adjacency(1|<grouping factor>)), and when the covStruct syntax is used to provide a fixed precision matrix (see pedigree for an example). A non-default choice of fitting algorithm can be selected in this and other models by using spaMM.options(sparse_precision= <TRUE|FALSE>) with often poor results. A precision matrix is meaningful for a Gaussian random effect, but beyond this the algorithm works for HGLMs, i.e. the model may include another random effect with non-Gaussian distribution.

Arguments