determine_npc: Determine the number of FPCs based on the share of explained...
In julia-wrobel/registr: Curve Registration for Exponential Family Functional Data

This internal function is called in gfpca_twoStep, fpca_gauss and bfpca to determine the number of functional principal components based on their share of explained variance.

1	determine_npc(evalues, npc_criterion)

evalues

Vector of estimated variances of the FPC scores.

npc_criterion

Either (i) a share between 0 and 1, or (ii) a vector with two elements for the targeted explained share of variance and a cut-off scree plot criterion, both between 0 and 1. For the latter, e.g., npc_criterion = c(0.9,0.02) tries to choose a number of FPCs that explains at least 90% of variation, but only includes FPCs that explain at least 2% of variation (even if this means 90% explained variation is not reached).

Integer for the number of fucntional principal components

julia-wrobel/registr documentation built on Jan. 16, 2022, 2:51 a.m.