Module fpca_variants 

Specialized functional-PCA variants.

This module collects the specialized FPCA / cross-covariance tools that the R ecosystems (fdapace, refund) expose and that fdars previously lacked:

• fpca_der — FPCA of curve derivatives (differentiate curves, then FPCA).
• fsvd — functional SVD / cross-FPCA between two paired functional samples.
• cross_covariance — the cross-covariance surface between two samples.
• dynamical_correlation — a scalar dynamical/functional correlation.
• ssvd — a sandwich-smoother / sparse-SVD FPCA path.

All entry points are additive and non-breaking: they reuse the dense FPCA engine (crate::regression::fdata_to_pc_1d) and the covariance/derivative helpers in crate::fdata / crate::covariance rather than introducing a new subsystem, and they add no new crate dependency. Every public function returns Result and validates its inputs up front (empty matrix, mismatched argument grids, mismatched sample sizes, ncomp out of range) rather than panicking. Outputs are numeric only — no plotting/rendering.

Structs

FsvdResult

Result of a functional SVD (fsvd) between two paired functional samples.

Functions

cross_covariance

Cross-covariance surface between two paired functional samples.

dynamical_correlation

Dynamical (functional) correlation between two paired functional samples.

fpca_der

FPCA of the derivatives of a functional sample.

fsvd

Functional SVD / cross-FPCA between two paired functional samples.

ssvd

Sandwich-smoother / sparse-SVD FPCA path.