ipw_select — IPW — Iterative Predictor Weighting

Group: Variable selector · Registry tolerance: 1e-06

Description

IPW-PLS iterative predictor weighting (§18 Phase 5t)

Registry note — R plsVarSel::ipw_pls iterative predictor weighting (RC filter, scale=TRUE, no.iter=3, IPW.threshold=0.01). Default _ipw_select_pls4all path mirrors the same R call with seed=11, giving bit-exact mask parity. The C++ top-k iterative-score kernel is opt-in via legacy=True.

Parameters

Name	Type	Default	Notes
top_k	int	required	current public binding signature
n_components	int	2	current public binding signature
n_iterations	int	20	current public binding signature
damping	float	0.5	current public binding signature
weight_floor	float	1e-06	current public binding signature
n_folds	int	3	current public binding signature
seed	int	0	current public binding signature

Explanations

Bibliographic source

Forina, M., Casolino, C. & Pizarro Millán, C. (1999). Iterative predictor weighting (IPW) PLS: a technique for the elimination of useless predictors in regression problems. Journal of Chemometrics 13(2), 165–184. https://doi.org/10.1002/(SICI)1099-128X(199903/04)13:2<165::AID-CEM535>3.0.CO;2-Y

Mathematical principle

IPW iteratively re-weights features in \(\mathbf{X}\) by their importance, refits PLS on the re-weighted data, and tracks the score path. Weights are derived from coefficient magnitude after each fit; the iteration converges to a stable importance ranking.

Compared to single-fit coefficient ranking, IPW’s iterative refinement gives more stable rankings when the calibration set is small or noisy. Exposes both the score path (for diagnostic) and the weight path (for interpretation).

Appropriate uses

Iteratively reweighting predictors from PLS coefficients to concentrate a model on consistently influential wavelengths.

Limits and validation

The update is heuristic and can amplify early ranking errors; convergence does not establish causal relevance.

Implementation

n4m_feature_selection_ipw_select.

Sources and provenance

Current implementation: cpp/src/core/ipw_selection.cpp.

API and bindings

C ABI (ABI 2): n4m_feature_selection_ipw_select. Use the linked public header for the exact signature, configuration, and result handles.

Python (verified public re-export):

from n4m.feature_selection.wrapper import IPW

Source signature: IPW(top_k: int, *, n_components: int = 2, n_iterations: int = 20, damping: float = 0.5, weight_floor: float = 1e-06, n_folds: int = 3, seed: int = 0) (n4m/_impl/selection.py).

R (source-verified): ipw_select(X, Y, n_components, n_iterations = 10L, top_k = 10L, damping = 0.5, weight_floor = 1e-6).

The source signature has additional required inputs, so no example call is fabricated.

MATLAB / Octave (source-verified): ipw_select(X, Y, n_components, n_iterations, top_k, damping, weight_floor).

The source signature has additional required inputs, so no example call is fabricated.

Registry parity references 📐

• 📐 ref.r_plsvarsel (R · r) — plsVarSel 0.10.0 · strict (rmse_rel ≤ 1e-06) — R plsVarSel::ipw_pls — iterative predictor weighting with the RC filter.

Benchmarks

Archived measurement identity. Backend labels in this table are the raw IDs recorded when the benchmark ran (including historical pls4all.* IDs). They preserve measurement provenance and do not describe a current public Python, R, or MATLAB binding; use the source-verified API and bindings section above for current entry points.

Adaptive wall-clock per cell measured against full_matrix.csv. Only backends that implement this method are listed; libraries without the method are omitted.

Verdict  ·  ✓ ref / ≈ ref / ~ shape mark a reference-gate pass at strict / relaxed / qualitative tolerance  ·  ✓ bind = archived binding-harness result agrees with the C++ baseline  ·  ⇄ cross-check = documented by-design selector/RNG/model, noncanonical API/facade convention, or secondary oracle  ·  ✗ divergent  ·  ⚠ error  ·  — not run. The fastest backend per column is marked 🏆.

Reference gate: strict — numeric equivalence (rmse_rel_tol ≤ 1e-06).

Rows tagged with 📐 are the canonical parity references for this method (declared in parity_timing.registry). C++ and external rows show reference parity; archived language-harness rows show binding parity against the C++ backend. Hover the icon for role and tolerance band.

1 thread

Backend	Parity	200×40 (ms)
C++ native · libn4m
pls4all.cpp.blas+omp	✓ J 1.00	360.4 ms
Python · archived pls4all benchmark
pls4all.python	✓ J 1.00	366.9 ms
pls4all.sklearn	⇄ J 0.75	1.81 ms🏆
R · archived pls4all benchmark
pls4all.R	⇄ J 0.75	4.86 ms
pls4all.R.formula	⇄ J 0.75	5.51 ms
pls4all.R.mdatools	⇄ J 0.75	5.15 ms
pls4all.R.pls	⇄ J 0.75	5.86 ms
R · external
📐ref.r_plsvarsel	source	22.0 ms

3 threads

Backend	Parity	200×40 (ms)
C++ native · libn4m
pls4all.cpp.blas+omp	✓ J 1.00	368.1 ms
Python · archived pls4all benchmark
pls4all.python	✓ J 1.00	358.9 ms
pls4all.sklearn	⇄ J 0.75	1.82 ms🏆
R · archived pls4all benchmark
pls4all.R	⇄ J 0.75	4.32 ms
pls4all.R.formula	⇄ J 0.75	5.47 ms
pls4all.R.mdatools	⇄ J 0.75	5.38 ms
pls4all.R.pls	⇄ J 0.75	5.37 ms
R · external
📐ref.r_plsvarsel	source	20.7 ms

10 threads

Backend	Parity	200×40 (ms)
C++ native · libn4m
pls4all.cpp.blas+omp	✓ J 1.00	392.3 ms
Python · archived pls4all benchmark
pls4all.python	✓ J 1.00	392.5 ms
pls4all.sklearn	⇄ J 0.75	1.86 ms🏆
R · archived pls4all benchmark
pls4all.R	⇄ J 0.75	4.39 ms
pls4all.R.formula	⇄ J 0.75	5.60 ms
pls4all.R.mdatools	⇄ J 0.75	5.39 ms
pls4all.R.pls	⇄ J 0.75	5.70 ms
R · external
📐ref.r_plsvarsel	source	26.2 ms

---

See also: benchmark overview · methods index · interactive dashboard