Module andrews 

Andrews curves transformation for multivariate and functional data.

Andrews curves map p-dimensional observations to functional curves using a Fourier-like bijection:

f_x(t) = x₁/√2 + x₂·sin(t) + x₃·cos(t) + x₄·sin(2t) + x₅·cos(2t) + …

for t ∈ [−π, π].

This provides a dimension-reduction visualization: each multivariate observation becomes a curve, and observations that are close in the original space produce curves that are close pointwise.

Structs

AndrewsLoadings

Andrews loadings: transform principal component loadings to Andrews curves.

AndrewsResult

Result of Andrews curves transformation.

Functions

andrews_loadings

Transform FPCA loadings to Andrews curves.

andrews_transform

Transform multivariate observations to Andrews curves.