aom_ridge_pls_superblock — n4m.compose.aom_superblock.aom_ridge_pls_superblock

Namespace: n4m.compose.aom_superblock · Fully-qualified: n4m.compose.aom_superblock.aom_ridge_pls_superblock · Catalog id: aom_pop.aom_ridge_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_ridge_pls_superblock

Signature: aom_ridge_pls_superblock(X, y, *, operators = None, n_components: int = 2, pls_components: Sequence[int] | None = None, ridge_lambda: float | None = None, ridge_lambdas: Sequence[float] = (0.0, 0.1, 1.0, 10.0), cv: int = 5, fold_ids = None, block_scaling: str = 'rms', center_x: bool = True)

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

ridge_lambda

float | None

None

ridge_lambdas

Sequence[float]

(0.0, 0.1, 1.0, 10.0)

cv

int

5

fold_ids

—

None

block_scaling

str

'rms'

center_x

bool

True

Explanations

Bibliographic source

No canonical publication defines this combined superblock head. 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

Build \(Z=[Z_1|\cdots|Z_B]\in\mathbb{R}^{n\times Bp}\) from strict operator outputs, center columns, and optionally RMS-scale each block. Ridge-PLS is then fitted on \(Z\) while CV evaluates the declared component-count and \(\lambda\) grids; the winning preprocessing and head are refit on all training rows.

Appropriate uses

Correlated multi-branch spectral representations where a linear, coefficient-exporting head is required.

Limits and validation

The additional Ridge penalty and number of components must be selected fold-locally without using a test set. Only the declared strict-linear branches support raw-space replay; block scaling changes the penalty geometry.

Implementation

n4m.compose.aom_superblock.aom_ridge_pls_superblock and AOMRidgePLSSuperblock; 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 Ridge-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 component count and Ridge-PLS lambda by train-fold CV, fits the native ridge-augmented SIMPLS, 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, MKL/kernel modes and dataset/source routing.

Timing benchmark: benchmarks/cross_binding/bench_aom_ridge_pls_superblock_timing.py

See also: methods index.