aom_pls_superblock — n4m.compose.aom_superblock.aom_pls_superblock

Namespace: n4m.compose.aom_superblock · Fully-qualified: n4m.compose.aom_superblock.aom_pls_superblock · Catalog id: aom_pop.aom_pls_superblock

API surface

C ABI: no standalone exported symbol is declared for this method.

Python (verified public re-export): from n4m.compose.aom_superblock import aom_pls_superblock

Signature: aom_pls_superblock(X, y, *, operators = None, n_components: int = 2, pls_components: Sequence[int] | None = None, cv: int = 5, fold_ids = None, block_scaling: str = 'rms', center_x: bool = True, center_y: bool = True, cuda_pls_parallel_folds: bool | None = None, cuda_pls_min_device_features: int | None = None, cuda_pls_many_batched: bool | None = None)

R: no current source-verified entry point was found for this catalog method.

MATLAB / Octave: no current source-verified entry point was found for this catalog method.

Parameters

Name	Type	Default
X	—	required
y	—	required
operators	—	None
n_components	int	2
pls_components	Sequence[int] | None	None
cv	int	5
fold_ids	—	None
block_scaling	str	'rms'
center_x	bool	True
center_y	bool	True
cuda_pls_parallel_folds	bool | None	None
cuda_pls_min_device_features	int | None	None
cuda_pls_many_batched	bool | None	None

Explanations

Bibliographic source

No publication specifies this ABI-2 superblock wrapper; it composes the AOM-PLS family. Beurier, G. et al. (2026). AOM-PLS / POP-PLS paper companion, arXiv:2605.13587, https://arxiv.org/abs/2605.13587. The product variants below are implementation-specific extensions; the paper does not by itself specify their ABI-2 orchestration.

Mathematical principle

For \(B\) declared operators, form outputs \(Z_b\in\mathbb{R}^{n\times p}\) and concatenate \(Z=[Z_1|\cdots|Z_B]\in\mathbb{R}^{n\times Bp}\). Columns are centered and each block can be RMS-scaled before PLS is fitted on \(Z\); the component-count grid is evaluated fold-locally and the selected head is refit on the full training rows.

Appropriate uses

Multi-block spectroscopy where distinct preprocessing branches are intentionally retained as model inputs.

Limits and validation

Block construction changes the feature geometry and can overweight blocks with more columns. Centering, RMS scaling, and component selection must be fold-local; raw-space fold-back is only valid for the declared strict-linear branches.

Implementation

n4m.compose.aom_superblock.aom_pls_superblock and AOMPLSSuperblock; no standalone C ABI symbol.

Sources and provenance

https://github.com/GBeurier/nirs4all-methods/blob/main/bindings/python/src/n4m/_impl/native.py; https://github.com/GBeurier/nirs4all-methods/blob/main/bindings/python/src/n4m/compose/aom_superblock.py

Catalog note

Native AOM-PLS superblock over strict-linear single-operator AOM views (core fit_aom_superblock). It concatenates the operator outputs with train-fold centering and optional block RMS scaling, selects the PLS component count by train-fold CV, fits the native moment/materialized PLS final fit, and folds the final superblock coefficients back to original-input input_coefficients plus intercept (affine N4MM state). It excludes row-reference-dependent preprocessing, nonlinear lifts and dataset/source routing.

Timing benchmark: benchmarks/cross_binding/bench_aom_pls_superblock_timing.py

See also: methods index.