iriv_select — IRIV — Iteratively Retaining Informative Variables

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

Description

Iteratively Retains Informative Variables (Phase 51)

Registry note — NumPy port of libPLS iriv (Yun 2014). Mann-Whitney U test via scipy.stats.mannwhitneyu. Default _iriv_select_pls4all path invokes the same NumPy function with np.random.default_rng(seed), giving bit-exact mask parity. The C++ splitmix64 kernel is opt-in via legacy=True.

Parameters

Name	Type	Default	Notes
n_components	int	2	current public binding signature
max_rounds	int	5	current public binding signature
n_folds	int	5	current public binding signature
seed	int	0	current public binding signature
fold	int	3	registry benchmark cell value

Explanations

Bibliographic source

Yun, Y. H., Wang, W. T., Tan, M. L., Liang, Y. Z., Li, H. D., Cao, D. S., Lu, H. M. & Xu, Q. S. (2014). A strategy that iteratively retains informative variables for selecting optimal variable subset in multivariate calibration. Analytica Chimica Acta 807, 36–43. Verified primary link: https://doi.org/10.1016/j.aca.2013.11.032.

Mathematical principle

IRIV classifies each variable into four categories at each iteration: strongly informative, weakly informative, uninformative, interfering. The first two are retained, the last two eliminated. Iteration continues until no further interfering variables remain.

Categories are determined by Monte-Carlo subset analysis with a permutation-based reference distribution: each variable’s CV-RMSE contribution distribution is compared against the distribution under random subset inclusion. This four-way classification is more nuanced than single-threshold methods and tends to handle correlated predictors well (correlated features can both be ‘weakly informative’).

Appropriate uses

Iteratively comparing variable subsets to classify informative and interfering wavelengths.

Limits and validation

Repeated random subset tests are stochastic and computationally expensive; significance labels depend on the sampled combinations.

Implementation

n4m_feature_selection_iriv_select.

Sources and provenance

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

API and bindings

C ABI (ABI 2): n4m_feature_selection_iriv_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 IRIV

Source signature: IRIV(*, n_components: int = 2, max_rounds: int = 5, n_folds: int = 5, seed: int = 0) (n4m/_impl/selection.py).

R (source-verified): iriv_select(X, Y, n_components, max_rounds = 20L, seed = 0L).

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

MATLAB / Octave (source-verified): iriv_select(X, Y, n_components, max_rounds, seed).

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

Registry parity references 📐

• 📐 ref.python_iriv_numpy_port (python · python) — iriv_numpy_port 1.0.0 · strict (rmse_rel ≤ 1e-06) — NumPy port of libPLS iriv (Yun 2014). Mann-Whitney U test via scipy.stats.mannwhitneyu; binary-matrix sampler keyed to np.random.default_rng(seed) for bit-exact reproducibility.

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	80×25 (ms)
C++ native · libn4m
pls4all.cpp.blas+omp	✓ J 1.00	268.0 ms
Python · archived pls4all benchmark
pls4all.python	✓ J 1.00	270.2 ms
pls4all.sklearn	⇄ J 0.60	25.3 ms🏆
R · archived pls4all benchmark
pls4all.R	⇄ J 0.60	29.4 ms
pls4all.R.formula	⇄ J 0.60	30.0 ms
pls4all.R.mdatools	⇄ J 0.60	30.8 ms
pls4all.R.pls	⇄ J 0.60	30.5 ms
Python · external
📐ref.python_iriv_numpy_port	source	274.8 ms

3 threads

Backend	Parity	80×25 (ms)
C++ native · libn4m
pls4all.cpp.blas+omp	✓ J 1.00	273.2 ms
Python · archived pls4all benchmark
pls4all.python	✓ J 1.00	267.2 ms
pls4all.sklearn	⇄ J 0.60	25.1 ms🏆
R · archived pls4all benchmark
pls4all.R	⇄ J 0.60	29.8 ms
pls4all.R.formula	⇄ J 0.60	30.8 ms
pls4all.R.mdatools	⇄ J 0.60	30.6 ms
pls4all.R.pls	⇄ J 0.60	30.3 ms
Python · external
📐ref.python_iriv_numpy_port	source	275.1 ms

10 threads

Backend	Parity	80×25 (ms)
C++ native · libn4m
pls4all.cpp.blas+omp	✓ J 1.00	274.2 ms
Python · archived pls4all benchmark
pls4all.python	✓ J 1.00	275.4 ms
pls4all.sklearn	⇄ J 0.60	25.4 ms🏆
R · archived pls4all benchmark
pls4all.R	⇄ J 0.60	29.9 ms
pls4all.R.formula	⇄ J 0.60	30.7 ms
pls4all.R.mdatools	⇄ J 0.60	31.0 ms
pls4all.R.pls	⇄ J 0.60	30.7 ms
Python · external
📐ref.python_iriv_numpy_port	source	270.7 ms

---

See also: benchmark overview · methods index · interactive dashboard