← Back to the catalog
Private3 sources

CORN Eigenvector NIR

corn · NIR

CORN Eigenvector NIR. v2.0 standardized NIRS package: 3 spectral source(s), 4 declared target(s). Auto-generated from dataset_card.json (verify before publication).

nirv2corn
🔒
Private dataset. Full metadata and metrics are shown, but the bytes are not redistributed here — exporting the data requires a Dataverse token. The identity card carries no spectra, only descriptive statistics.
80
samples
2,100
wavelengths
3
sources
4
targets
0
metadata
NIR
family

Dataset property explorer

Mean profile risk0.42
Highest axisOutliers PCA · 1.00
Diagnostics8
Sources profiled3
CORN Eigenvector NIR property profile0.250.50.751integritynoiseartefactsbaselinePCA outliersreferencerepeatabilitystructureCORN Eigenvector NIR profileintegrity: 0.00noise: 0.00artefacts: 0.00baseline: 1.00PCA outliers: 1.00reference: 0.36repeatability: 0.00structure: 1.00CORN Eigenvecto…0 center · 1 outer ring · outward = stronger anomaly / heterogeneity signal

Profile axes

Intégrité0.00
Artefacts locaux0.00
Bruit0.00
Outliers PCA1.00
Distance à la référence0.36
Répétabilité0.00
Baseline / forme1.00
Structure multi-régimes1.00
Diagnostic hypotheses00.250.50.751hypothesis scoreSpectre hors domaine valideSpectre hors domaine valide: 0.810.81Fond différentFond différent: 0.800.80Erreur calibration / référenc…Erreur calibration / référence blanche: 0.680.68Mélange feuille + fondMélange feuille + fond: 0.650.65Dataset multi-régimesDataset multi-régimes: 0.590.59Différence de sonde / géométr…Différence de sonde / géométrie: 0.530.53Signature VERA25-likeSignature VERA25-like: 0.500.50Erreur interpolation / réécha…Erreur interpolation / rééchantillonnage: 0.490.49
DiagnosticScoreForceSignauxInterprétation probable
Spectre hors domaine valideX10.81forteStructure PCA 1.00, Mahalanobis / T2 1.00, artefacts faibles 1.00Variété, espèce, lot ou condition différente mais physiquement plausible.
Fond différentX10.80fortePCA Q 1.00, Baseline/mean/area 1.00, Mahalanobis / T2 1.00Effet systématique du support, blanc/noir, transflectance ou environnement de mesure.
Erreur calibration / référence blancheX10.68moyennePCA Q 1.00, Baseline/mean/area 1.00, Mahalanobis / T2 1.00Décalage systématique entre campagnes, instruments ou référence blanche.
Mélange feuille + fondX10.65moyennePCA Q 1.00, Baseline/mean/area 1.00, Mahalanobis / T2 1.00Couverture partielle du spot; contribution du fond ou du support.
Dataset multi-régimesX10.59moyenneStructure PCA 1.00, PCA Q 1.00, Mahalanobis / T2 1.00Mélange de campagnes, opérateurs, lots, setups ou sous-populations spectrales.
Différence de sonde / géométrieX10.53moyennePCA Q 1.00, Baseline/mean/area 1.00, Mahalanobis / T2 1.00Modification de l'illumination, collecte, angle ou distance sonde-échantillon.
Signature VERA25-likeX10.50moyennePCA Q 1.00, Mahalanobis / T2 1.00, RMS/SAM référence 0.37Combinaison possible changement de sonde + splice, amplifiée par géométrie, fond ou calibration.
Erreur interpolation / rééchantillonnageX30.49moyennePCA Q 1.00, SNR normal/élevé 1.00, Noise RMS faible 1.00Artefacts numériques ou traitement spectral incorrect.

Spectral sources

m5spec

X1 · NIR · m5
m5spec spectra0.000.250.500.751.001,0001,5002,0002,500q05-q95 envelopeq25-q75 envelopemedian spectrummedianq25–q75q05–q95wavelength / nm1,100nm — median 0.04563 (q25–q75 0.04044–0.04933)1,110nm — median 0.04535 (q25–q75 0.04012–0.04902)1,120nm — median 0.04671 (q25–q75 0.04129–0.05039)1,130nm — median 0.05228 (q25–q75 0.04733–0.05715)1,140nm — median 0.06531 (q25–q75 0.06076–0.07147)1,150nm — median 0.08195 (q25–q75 0.0778–0.09053)1,160nm — median 0.09898 (q25–q75 0.09381–0.1096)1,170nm — median 0.1122 (q25–q75 0.1067–0.1241)1,180nm — median 0.1232 (q25–q75 0.1178–0.1365)1,190nm — median 0.1335 (q25–q75 0.1276–0.1471)1,200nm — median 0.1384 (q25–q75 0.1321–0.1518)1,210nm — median 0.1361 (q25–q75 0.1299–0.1492)1,220nm — median 0.127 (q25–q75 0.1212–0.1396)1,230nm — median 0.1167 (q25–q75 0.1113–0.1287)1,240nm — median 0.1094 (q25–q75 0.1042–0.1207)1,250nm — median 0.1051 (q25–q75 0.1–0.116)1,260nm — median 0.1026 (q25–q75 0.09758–0.1134)1,270nm — median 0.1006 (q25–q75 0.09559–0.1112)1,282nm — median 0.09746 (q25–q75 0.09263–0.108)1,292nm — median 0.09506 (q25–q75 0.09033–0.1054)1,302nm — median 0.09434 (q25–q75 0.08966–0.1046)1,312nm — median 0.09636 (q25–q75 0.0916–0.1068)1,322nm — median 0.1021 (q25–q75 0.09713–0.113)1,332nm — median 0.1123 (q25–q75 0.1068–0.1239)1,342nm — median 0.1263 (q25–q75 0.1199–0.1387)1,352nm — median 0.1419 (q25–q75 0.1345–0.1554)1,362nm — median 0.1566 (q25–q75 0.1482–0.1711)1,372nm — median 0.169 (q25–q75 0.1605–0.1844)1,382nm — median 0.1832 (q25–q75 0.1735–0.1989)1,392nm — median 0.2066 (q25–q75 0.1947–0.223)1,402nm — median 0.2422 (q25–q75 0.2277–0.2596)1,412nm — median 0.2869 (q25–q75 0.2701–0.3066)1,422nm — median 0.3313 (q25–q75 0.3125–0.3532)1,432nm — median 0.3655 (q25–q75 0.3449–0.3891)1,442nm — median 0.3849 (q25–q75 0.3634–0.4096)1,452nm — median 0.3935 (q25–q75 0.3717–0.4189)1,462nm — median 0.3969 (q25–q75 0.3748–0.4225)1,472nm — median 0.3964 (q25–q75 0.3743–0.4222)1,482nm — median 0.3923 (q25–q75 0.3705–0.4184)1,492nm — median 0.3864 (q25–q75 0.3651–0.4129)1,502nm — median 0.3795 (q25–q75 0.3587–0.406)1,512nm — median 0.3721 (q25–q75 0.3517–0.3982)1,522nm — median 0.3657 (q25–q75 0.3455–0.3909)1,532nm — median 0.3611 (q25–q75 0.3411–0.3856)1,542nm — median 0.3583 (q25–q75 0.3383–0.3826)1,552nm — median 0.3564 (q25–q75 0.3365–0.3806)1,562nm — median 0.3548 (q25–q75 0.3351–0.379)1,572nm — median 0.3524 (q25–q75 0.3328–0.3766)1,582nm — median 0.3481 (q25–q75 0.3287–0.3721)1,592nm — median 0.3411 (q25–q75 0.3219–0.3646)1,602nm — median 0.3315 (q25–q75 0.3128–0.3547)1,612nm — median 0.3212 (q25–q75 0.3028–0.3437)1,622nm — median 0.3112 (q25–q75 0.2931–0.3331)1,634nm — median 0.3003 (q25–q75 0.2826–0.3215)1,644nm — median 0.2926 (q25–q75 0.2752–0.3134)1,654nm — median 0.287 (q25–q75 0.2697–0.3075)1,664nm — median 0.2844 (q25–q75 0.2671–0.3049)1,674nm — median 0.2859 (q25–q75 0.2689–0.3069)1,684nm — median 0.2959 (q25–q75 0.2791–0.3183)1,694nm — median 0.3116 (q25–q75 0.2942–0.3355)1,704nm — median 0.3219 (q25–q75 0.3041–0.3465)1,714nm — median 0.3246 (q25–q75 0.3065–0.3498)1,724nm — median 0.3269 (q25–q75 0.3085–0.3522)1,734nm — median 0.3285 (q25–q75 0.3102–0.3536)1,744nm — median 0.3321 (q25–q75 0.313–0.3567)1,754nm — median 0.3359 (q25–q75 0.3164–0.3605)1,764nm — median 0.3373 (q25–q75 0.3177–0.3618)1,774nm — median 0.3378 (q25–q75 0.318–0.3617)1,784nm — median 0.3349 (q25–q75 0.3153–0.3583)1,794nm — median 0.3277 (q25–q75 0.3087–0.3509)1,804nm — median 0.3198 (q25–q75 0.3014–0.3429)1,814nm — median 0.3138 (q25–q75 0.2959–0.3368)1,824nm — median 0.3087 (q25–q75 0.2911–0.3312)1,834nm — median 0.3023 (q25–q75 0.2849–0.3239)1,844nm — median 0.2954 (q25–q75 0.2782–0.3162)1,854nm — median 0.2925 (q25–q75 0.2754–0.3132)1,864nm — median 0.2975 (q25–q75 0.28–0.3186)1,874nm — median 0.314 (q25–q75 0.2952–0.3351)1,884nm — median 0.3476 (q25–q75 0.3265–0.3697)1,894nm — median 0.4073 (q25–q75 0.3824–0.4318)1,904nm — median 0.4808 (q25–q75 0.4545–0.5086)1,914nm — median 0.5427 (q25–q75 0.5156–0.5712)1,924nm — median 0.5795 (q25–q75 0.5525–0.6108)1,934nm — median 0.5898 (q25–q75 0.5628–0.623)1,944nm — median 0.5778 (q25–q75 0.5506–0.611)1,954nm — median 0.5576 (q25–q75 0.5299–0.5898)1,964nm — median 0.5364 (q25–q75 0.5088–0.5678)1,976nm — median 0.5118 (q25–q75 0.4834–0.5422)1,986nm — median 0.4919 (q25–q75 0.4643–0.5224)1,996nm — median 0.4754 (q25–q75 0.4486–0.5058)2,006nm — median 0.4656 (q25–q75 0.4392–0.4959)2,016nm — median 0.4651 (q25–q75 0.4387–0.4953)2,026nm — median 0.4744 (q25–q75 0.4476–0.5048)2,036nm — median 0.4915 (q25–q75 0.4639–0.5217)2,046nm — median 0.5116 (q25–q75 0.4833–0.5415)2,056nm — median 0.5306 (q25–q75 0.5023–0.5599)2,066nm — median 0.5455 (q25–q75 0.5177–0.5745)2,076nm — median 0.5567 (q25–q75 0.53–0.5857)2,086nm — median 0.5649 (q25–q75 0.5388–0.5938)2,096nm — median 0.5698 (q25–q75 0.5441–0.5987)2,106nm — median 0.5712 (q25–q75 0.5456–0.6002)2,116nm — median 0.5695 (q25–q75 0.5436–0.5984)2,126nm — median 0.5646 (q25–q75 0.5383–0.5935)2,136nm — median 0.5569 (q25–q75 0.5299–0.5857)2,146nm — median 0.5473 (q25–q75 0.5191–0.576)2,156nm — median 0.536 (q25–q75 0.5072–0.5647)2,166nm — median 0.5246 (q25–q75 0.4965–0.554)2,176nm — median 0.5132 (q25–q75 0.4859–0.5428)2,186nm — median 0.5015 (q25–q75 0.475–0.5313)2,196nm — median 0.4902 (q25–q75 0.4644–0.5197)2,206nm — median 0.4801 (q25–q75 0.4551–0.5097)2,216nm — median 0.4723 (q25–q75 0.4476–0.5016)2,226nm — median 0.4686 (q25–q75 0.444–0.4976)2,236nm — median 0.4735 (q25–q75 0.4485–0.5019)2,246nm — median 0.4903 (q25–q75 0.4643–0.5193)2,256nm — median 0.5176 (q25–q75 0.4909–0.5479)2,266nm — median 0.5481 (q25–q75 0.5203–0.579)2,276nm — median 0.5757 (q25–q75 0.5485–0.6087)2,286nm — median 0.5903 (q25–q75 0.5634–0.6247)2,296nm — median 0.5893 (q25–q75 0.5622–0.624)2,306nm — median 0.5996 (q25–q75 0.5722–0.6342)2,316nm — median 0.6038 (q25–q75 0.5761–0.6389)2,328nm — median 0.6011 (q25–q75 0.5732–0.6351)2,338nm — median 0.597 (q25–q75 0.5691–0.6311)2,348nm — median 0.5936 (q25–q75 0.5659–0.6278)2,358nm — median 0.5888 (q25–q75 0.5609–0.6225)2,368nm — median 0.5884 (q25–q75 0.5604–0.6222)2,378nm — median 0.5879 (q25–q75 0.5598–0.6218)2,388nm — median 0.5888 (q25–q75 0.5606–0.6229)2,398nm — median 0.5956 (q25–q75 0.5675–0.6302)2,408nm — median 0.6099 (q25–q75 0.5818–0.6448)2,418nm — median 0.628 (q25–q75 0.5997–0.663)2,428nm — median 0.6467 (q25–q75 0.6186–0.6816)2,438nm — median 0.667 (q25–q75 0.639–0.7018)2,448nm — median 0.6879 (q25–q75 0.6602–0.7226)2,458nm — median 0.7067 (q25–q75 0.6788–0.741)2,468nm — median 0.7237 (q25–q75 0.6958–0.7577)2,478nm — median 0.736 (q25–q75 0.7086–0.7702)2,488nm — median 0.7416 (q25–q75 0.7146–0.7756)2,498nm — median 0.7412 (q25–q75 0.7147–0.7751)

Sampling

Wavelengths700
Axis range1,100–2,498 nm
Mean spacing2 nm
Griduniform
Observations80

Signal & quality

Value range0.0317 – 0.878
Mean range0.0454 – 0.748
Mean level0.389
Area543.8
PTP0.7028
Noise RMS8.1261e-05
SNR4.8e+03
SNR dB7e+01 dB
Dynamic range0.703
Smoothness0.0003185
Saturated0.0%
X-outliers28

Integrity & artefacts

NaN ratio0.00%
Inf count0
Zero ratio0.00%
Spike count2
Spike rate0.00%
Jump count0
Jump rate0.00%
Clip fraction0.00%

Shape & reference

Baseline slope0.5948
Curvature RMS0.00031958
D1 RMS0.0033338
RMS to mean0.02317
RMS p950.065148
SAM to mean0.0086784
SAM p950.02101
Affine offset p950.017854
Affine gain p95 Δ0.12642
Affine residual p950.0046742
Xcorr lag p950

Outliers & repeatability

PCA Q p95/median10
Hotelling T2 p95/median8
Mahalanobis H p95/median2.8
Repeat groups0

Dimensionality (PCA)

Effective rank1.1
PCs → 95% var1
PCs → 99% var1
Top-10 cum. var100.0%
Computed metric scores 29worst 1.00
FamilleMétrique calculéeValeurScoreNiveauInterprétation datasetCauses typiquesCalcul / scoring
Intégrité des donnéesNaN ratiointegrity.nan_ratio0%0.00faibleSpectre completErreur acquisition/exportcount(isnan(X)) / X.sizealert = min(1, nan_ratio / 0.05)
Intégrité des donnéesInf countintegrity.inf_count00.00faibleNormalCalculs invalidescount(isinf(X))alert = min(1, inf_count / 1)
Intégrité des donnéesZero ratiointegrity.zero_ratio0%0.00faibleNormalExport, saturationcount(X == 0) / count(finite X)alert = min(1, zero_ratio / 0.05)
Amplitude globaleMean reflectanceamplitude.mean_reflectance0.389011.00fortValeur atypique: Trop clair / fond visible ou Trop sombreFond, géométriemean(X finite)alert reuses baseline/shape drift because absolute reflectance ranges are technology-dependent
Amplitude globaleArea under curveamplitude.area_under_curve543.821.00fortValeur atypique: Différence d'éclairement ou NormalDistance sondetrapezoid(mean_spectrum, spectral_axis)alert reuses baseline/shape drift because area scale depends on axis and units
Amplitude globalePeak-to-peak (PTP)amplitude.peak_to_peak0.702850.00faibleVariabilité forteSaturationmax(mean_spectrum) - min(mean_spectrum)alert increases when dynamic range is abnormally flat
Amplitude globaleVarianceamplitude.variance0.0351840.00faibleNormal ou hétérogèneMauvais contactvar(X finite)alert increases when variance/dynamic range is abnormally flat
BruitNoise RMSnoise.noise_rms8.1261e-050.00faibleStableLampe, détecteurmedian MAD(second derivative) * 1.4826 / sqrt(6)alert = noise_rms / signal_scale, saturated at 5%
BruitSNRnoise.snr4787.10.00faibleBon signalAcquisitionmean(abs(X)) / noise_rmsalert decreases with SNR dB; >=40 dB is treated as low alert
BruitBandwise SNRnoise.bandwise_snr_min2614.60.00faibleZone fiableDétecteurmin(abs(mean_spectrum) / local second-derivative noise)alert decreases with worst-band SNR dB; >=35 dB is treated as low alert
Artefacts locauxSpike countartefacts.spike_count20.00faibleSpectre propreCosmic rays, splicecount robust outliers in second derivativealert follows spike_rate, saturated at 1%
Artefacts locauxSpike rateartefacts.spike_rate0.00358%0.00faibleNormalInterpolationspike_count / (n_samples * (n_features - 2))alert = min(1, spike_rate / 0.01)
Artefacts locauxJump countartefacts.jump_count00.00faibleContinuSplicecount robust outliers in first derivativealert follows jump_rate, saturated at 1%
Artefacts locauxJump rateartefacts.jump_rate0%0.00faibleNormalCalibrationjump_count / (n_samples * (n_features - 1))alert = min(1, jump_rate / 0.01)
Artefacts locauxClip fractionartefacts.clip_fraction0.00357%0.00faibleNormalDétecteur saturéfraction of finite cells equal to repeated min/max extremaalert = min(1, clip_fraction / 0.01)
Forme spectraleBaseline slopeshape.baseline_slope0.59481.00fortDériveÉclairementlinear slope of mean_spectrum over normalized axisalert = abs(slope / signal_scale), saturated at 0.5
Forme spectraleCurvature RMSshape.curvature_rms0.000319580.05faibleLisseFond, splicemedian RMS(second derivative per spectrum)alert = curvature_rms / signal_scale, saturated at 1%
Forme spectraleD1 RMSshape.d1_rms0.00333380.09faiblePlatBiologie ou artefactmedian RMS(first derivative per spectrum)alert = d1_rms / signal_scale, saturated at 5%
Outliers multivariésPCA Q (SPE)outliers.pca_q_ratio10.3681.00fortSpectre atypiqueArtefact, mélangep95(Q/SPE residual) / median(Q/SPE residual)alert = min(1, pca_q_ratio / 8)
Outliers multivariésHotelling T²outliers.hotelling_t2_ratio7.97591.00fortExtrême mais cohérentVariabilité naturellep95(Hotelling T2) / median(Hotelling T2)alert = min(1, hotelling_t2_ratio / 8)
Outliers multivariésMahalanobis Houtliers.mahalanobis_h_ratio2.82410.71moyenOutlier globalDomaine différentp95(sqrt(T2)) / median(sqrt(T2))alert = min(1, mahalanobis_h_ratio / 4)
Comparaison à référenceRMS to mean spectrumreference.rms_to_mean_spectrum_p950.0651480.37faibleTypiqueDomain shiftp95 RMS distance to dataset mean spectrumalert = RMS_p95 / signal_scale, saturated at 25%
Comparaison à référenceSpectral Angle Mapper (SAM)reference.sam_to_mean_spectrum_p950.021010.06faibleSimilaireFond, géométriep95 spectral angle to dataset mean spectrumalert = min(1, SAM_p95 / 0.35 rad)
RépétabilitéRMS intra-IDrepeatability.rms_intra_id—0.00faibleStablePositionnementmedian RMS distance to repeated-sample centroidalert = RMS_intra_ID / signal_scale, saturated at 10%
RépétabilitéSAM intra-IDrepeatability.sam_intra_id—0.00faibleStableAcquisitionmedian SAM to repeated-sample centroidalert = min(1, SAM_intra_ID / 0.15 rad)
RépétabilitéCV intra-IDrepeatability.cv_intra_id—0.00faibleStableOpérateurmedian within-ID band CValert = min(1, CV_intra_ID / 0.25)
Structure du datasetPCA score densitystructure.pca_score_density11.3711.00fortSous-populationsLots différents1 / median kNN distance in PCA score spacealert follows density_cv/profile structure complexity, not raw density alone
Structure du datasetLocal Outlier Factor (LOF)structure.local_outlier_factor_p954.02551.00fortSpectre isoléCas raresp95 approximate LOF from PCA-score kNN distancesalert = min(1, max(0, LOF_p95 - 1) / 2)
Structure du datasetIsolation Forest scorestructure.isolation_forest_score_p950.629241.00fortSpectre atypiqueDiverses causesp95 IsolationForest anomaly score on PCA scoresalert follows structure complexity; raw score is implementation-dependent
X1 PCA score plot-2024-0.20.00.20.4PC1 -0.3914 · PC2 -0.01728PC1 -0.4439 · PC2 0.01701PC1 -0.726 · PC2 0.06928PC1 -0.9611 · PC2 0.013PC1 0.6062 · PC2 0.05031PC1 0.556 · PC2 -0.03006PC1 0.6132 · PC2 0.001193PC1 0.799 · PC2 0.02895PC1 -0.6973 · PC2 0.05965PC1 -0.3444 · PC2 0.0002117PC1 1.922 · PC2 0.008436PC1 1.707 · PC2 -0.075PC1 -0.4616 · PC2 -0.06275PC1 -0.2251 · PC2 -0.02696PC1 1.366 · PC2 -0.05729PC1 1.787 · PC2 -0.07572PC1 0.2806 · PC2 0.0513PC1 0.8364 · PC2 -0.06455PC1 0.4684 · PC2 0.04947PC1 0.9028 · PC2 -0.03699PC1 0.3378 · PC2 -0.06864PC1 0.9091 · PC2 -0.06124PC1 0.9012 · PC2 -0.03112PC1 0.8992 · PC2 -0.03705PC1 0.5851 · PC2 0.07579PC1 0.1739 · PC2 -0.01558PC1 1.589 · PC2 -0.08793PC1 1.582 · PC2 -0.1212PC1 1.396 · PC2 -0.0517PC1 -1.229 · PC2 -0.01023PC1 -1.105 · PC2 -0.03625PC1 -0.2913 · PC2 0.01301PC1 -0.1389 · PC2 0.008601PC1 0.7633 · PC2 -0.09945PC1 0.9264 · PC2 0.006585PC1 -1.034 · PC2 0.003073PC1 -1.165 · PC2 0.03106PC1 -0.808 · PC2 0.03479PC1 -1.073 · PC2 0.01861PC1 -1.386 · PC2 -0.03336PC1 -1.252 · PC2 -0.03171PC1 -1.416 · PC2 0.07361PC1 -1.314 · PC2 -0.002412PC1 -0.4747 · PC2 0.04877PC1 -0.6895 · PC2 0.05791PC1 0.004708 · PC2 0.06275PC1 0.6339 · PC2 0.07589PC1 -0.2135 · PC2 0.09507PC1 -0.4164 · PC2 0.005199PC1 -0.5302 · PC2 0.02846PC1 -0.1001 · PC2 0.01596PC1 -0.2033 · PC2 -0.008925PC1 0.2178 · PC2 -0.004915PC1 -1.628 · PC2 0.01221PC1 -1.719 · PC2 0.02126PC1 -0.736 · PC2 -0.1416PC1 0.2514 · PC2 -0.1796PC1 0.07178 · PC2 -0.05476PC1 -0.06835 · PC2 -0.08991PC1 0.1357 · PC2 -0.03619PC1 -0.6067 · PC2 -0.01191PC1 -0.04283 · PC2 -0.01491PC1 0.4118 · PC2 -0.01878PC1 -0.3661 · PC2 -0.09052PC1 -0.2245 · PC2 -0.05879PC1 -1.122 · PC2 -0.03097PC1 0.03542 · PC2 -0.06283PC1 -0.7567 · PC2 -0.06071PC1 0.03578 · PC2 -0.07088PC1 -0.7377 · PC2 0.03916PC1 0.009705 · PC2 -0.1502PC1 0.1067 · PC2 -0.1939PC1 -0.6006 · PC2 0.06807PC1 -1.318 · PC2 0.1165PC1 2.904 · PC2 0.2057PC1 -0.1317 · PC2 0.1079PC1 2.316 · PC2 0.2654PC1 0.1615 · PC2 0.1386PC1 0.05254 · PC2 0.231PC1 -0.1053 · PC2 0.2049PC1 (99.1%)PC2 (0.8%)80 scores
PCA explained variance0%25%50%75%100%PC1: 99.1% (cumulative 99.1%)1PC2: 0.8% (cumulative 99.8%)2PC3: 0.1% (cumulative 99.9%)3PC4: 0.0% (cumulative 99.9%)4PC5: 0.0% (cumulative 100.0%)5PC6: 0.0% (cumulative 100.0%)6PC7: 0.0% (cumulative 100.0%)7PC8: 0.0% (cumulative 100.0%)8PC9: 0.0% (cumulative 100.0%)9PC10: 0.0% (cumulative 100.0%)10cumulative explained variancePC variancecumulativeprincipal component · cumulative (dashed)
X-Y spectral correlation 4
X1 · Moisture spectral correlation-1-0.500.51absolute correlation envelopesigned correlationabsolute correlation1,0001,5002,0002,500|r|signed raxis · Pearson correlation scale
X1 · Oil spectral correlation-1-0.500.51absolute correlation envelopesigned correlationabsolute correlation1,0001,5002,0002,500|r|signed raxis · Pearson correlation scale
X1 · Protein spectral correlation-1-0.500.51absolute correlation envelopesigned correlationabsolute correlation1,0001,5002,0002,500|r|signed raxis · Pearson correlation scale
Targetmax |r|axis @ maxmean |r||r| ≥ .5
Moisture0.6712,4780.58995.7%
Oil0.3082,3060.2440.0%
Protein0.3432,3040.3020.0%
Starch0.1591,4240.1320.0%

mp5spec

X2 · NIR · mp5
mp5spec spectra-0.250.000.250.500.751.001,0001,5002,0002,500q05-q95 envelopeq25-q75 envelopemedian spectrummedianq25–q75q05–q95wavelength / nm1,100nm — median -0.01077 (q25–q75 -0.01397–-0.006702)1,110nm — median -0.01078 (q25–q75 -0.014–-0.00666)1,120nm — median -0.009283 (q25–q75 -0.01248–-0.005065)1,130nm — median -0.003323 (q25–q75 -0.006753–0.001382)1,140nm — median 0.01022 (q25–q75 0.005681–0.01536)1,150nm — median 0.02786 (q25–q75 0.02273–0.0345)1,160nm — median 0.04635 (q25–q75 0.03967–0.05431)1,170nm — median 0.06106 (q25–q75 0.0531–0.07016)1,180nm — median 0.07402 (q25–q75 0.06438–0.08345)1,190nm — median 0.08573 (q25–q75 0.07489–0.09531)1,200nm — median 0.09188 (q25–q75 0.08075–0.1016)1,210nm — median 0.09062 (q25–q75 0.07985–0.1002)1,220nm — median 0.08226 (q25–q75 0.07196–0.09137)1,230nm — median 0.07211 (q25–q75 0.06285–0.08078)1,240nm — median 0.0647 (q25–q75 0.05596–0.07299)1,250nm — median 0.06062 (q25–q75 0.05212–0.0686)1,260nm — median 0.05853 (q25–q75 0.05016–0.06635)1,270nm — median 0.05693 (q25–q75 0.04873–0.06471)1,282nm — median 0.05424 (q25–q75 0.04632–0.06196)1,292nm — median 0.05198 (q25–q75 0.04424–0.05958)1,302nm — median 0.0513 (q25–q75 0.04362–0.05889)1,312nm — median 0.05327 (q25–q75 0.04549–0.06107)1,322nm — median 0.05889 (q25–q75 0.05072–0.06709)1,332nm — median 0.06892 (q25–q75 0.06004–0.07804)1,342nm — median 0.08313 (q25–q75 0.0731–0.09316)1,352nm — median 0.09935 (q25–q75 0.08771–0.1098)1,362nm — median 0.1146 (q25–q75 0.1014–0.1253)1,372nm — median 0.1274 (q25–q75 0.113–0.1382)1,382nm — median 0.1406 (q25–q75 0.1255–0.152)1,392nm — median 0.1622 (q25–q75 0.146–0.175)1,402nm — median 0.1963 (q25–q75 0.1771–0.2098)1,412nm — median 0.2418 (q25–q75 0.2195–0.2561)1,422nm — median 0.2884 (q25–q75 0.2632–0.3029)1,432nm — median 0.3246 (q25–q75 0.2977–0.3396)1,442nm — median 0.345 (q25–q75 0.3177–0.3608)1,452nm — median 0.3543 (q25–q75 0.3269–0.3706)1,462nm — median 0.3582 (q25–q75 0.3306–0.3747)1,472nm — median 0.3583 (q25–q75 0.3306–0.3748)1,482nm — median 0.355 (q25–q75 0.3271–0.3713)1,492nm — median 0.3496 (q25–q75 0.3219–0.3658)1,502nm — median 0.3429 (q25–q75 0.3155–0.3592)1,512nm — median 0.3358 (q25–q75 0.3086–0.3521)1,522nm — median 0.3298 (q25–q75 0.3024–0.3457)1,532nm — median 0.3254 (q25–q75 0.2979–0.341)1,542nm — median 0.3227 (q25–q75 0.2952–0.3382)1,552nm — median 0.321 (q25–q75 0.2934–0.3364)1,562nm — median 0.3195 (q25–q75 0.292–0.3349)1,572nm — median 0.3173 (q25–q75 0.29–0.3327)1,582nm — median 0.3131 (q25–q75 0.2859–0.3282)1,592nm — median 0.3062 (q25–q75 0.2792–0.3209)1,602nm — median 0.2968 (q25–q75 0.27–0.311)1,612nm — median 0.2862 (q25–q75 0.2599–0.3)1,622nm — median 0.2758 (q25–q75 0.2501–0.2893)1,634nm — median 0.2641 (q25–q75 0.2395–0.2776)1,644nm — median 0.2558 (q25–q75 0.2319–0.2693)1,654nm — median 0.2495 (q25–q75 0.2264–0.263)1,664nm — median 0.2461 (q25–q75 0.224–0.26)1,674nm — median 0.2472 (q25–q75 0.2252–0.2615)1,684nm — median 0.257 (q25–q75 0.2348–0.2721)1,694nm — median 0.2737 (q25–q75 0.2507–0.2895)1,704nm — median 0.2848 (q25–q75 0.2613–0.3014)1,714nm — median 0.2882 (q25–q75 0.2647–0.3051)1,724nm — median 0.2906 (q25–q75 0.2668–0.3077)1,734nm — median 0.2916 (q25–q75 0.2674–0.3083)1,744nm — median 0.2944 (q25–q75 0.2699–0.3107)1,754nm — median 0.2983 (q25–q75 0.2736–0.3145)1,764nm — median 0.2999 (q25–q75 0.2752–0.3159)1,774nm — median 0.3 (q25–q75 0.2753–0.3157)1,784nm — median 0.297 (q25–q75 0.2723–0.3124)1,794nm — median 0.2897 (q25–q75 0.2655–0.3051)1,804nm — median 0.2817 (q25–q75 0.2578–0.297)1,814nm — median 0.2755 (q25–q75 0.2519–0.2907)1,824nm — median 0.2699 (q25–q75 0.2466–0.2851)1,834nm — median 0.2628 (q25–q75 0.2399–0.278)1,844nm — median 0.2551 (q25–q75 0.2327–0.2704)1,854nm — median 0.2517 (q25–q75 0.2294–0.2671)1,864nm — median 0.2561 (q25–q75 0.2334–0.2717)1,874nm — median 0.2713 (q25–q75 0.2474–0.2875)1,884nm — median 0.3038 (q25–q75 0.278–0.32)1,894nm — median 0.36 (q25–q75 0.3333–0.3788)1,904nm — median 0.4299 (q25–q75 0.4034–0.4538)1,914nm — median 0.4887 (q25–q75 0.463–0.5168)1,924nm — median 0.5249 (q25–q75 0.5008–0.5553)1,934nm — median 0.536 (q25–q75 0.5119–0.567)1,944nm — median 0.5242 (q25–q75 0.4992–0.5552)1,954nm — median 0.5039 (q25–q75 0.4781–0.5344)1,964nm — median 0.4837 (q25–q75 0.4577–0.5132)1,976nm — median 0.4612 (q25–q75 0.4346–0.489)1,986nm — median 0.443 (q25–q75 0.4157–0.4697)1,996nm — median 0.4281 (q25–q75 0.4005–0.4536)2,006nm — median 0.42 (q25–q75 0.3923–0.4448)2,016nm — median 0.4212 (q25–q75 0.3932–0.4452)2,026nm — median 0.4324 (q25–q75 0.404–0.4557)2,036nm — median 0.4509 (q25–q75 0.4221–0.474)2,046nm — median 0.4723 (q25–q75 0.4431–0.4955)2,056nm — median 0.4922 (q25–q75 0.4629–0.5157)2,066nm — median 0.5084 (q25–q75 0.4789–0.5325)2,076nm — median 0.5211 (q25–q75 0.4915–0.5449)2,086nm — median 0.5302 (q25–q75 0.5003–0.5536)2,096nm — median 0.5359 (q25–q75 0.5057–0.5595)2,106nm — median 0.5378 (q25–q75 0.5074–0.5615)2,116nm — median 0.536 (q25–q75 0.5056–0.5596)2,126nm — median 0.5305 (q25–q75 0.4999–0.5539)2,136nm — median 0.5213 (q25–q75 0.4909–0.5449)2,146nm — median 0.5109 (q25–q75 0.4801–0.5349)2,156nm — median 0.4989 (q25–q75 0.468–0.5232)2,166nm — median 0.4873 (q25–q75 0.4564–0.5114)2,176nm — median 0.4756 (q25–q75 0.4454–0.4993)2,186nm — median 0.4634 (q25–q75 0.4336–0.4866)2,196nm — median 0.4521 (q25–q75 0.4223–0.4745)2,206nm — median 0.4423 (q25–q75 0.4124–0.4642)2,216nm — median 0.4345 (q25–q75 0.4047–0.4562)2,226nm — median 0.4307 (q25–q75 0.4008–0.4522)2,236nm — median 0.4349 (q25–q75 0.4052–0.4568)2,246nm — median 0.4515 (q25–q75 0.4219–0.4743)2,256nm — median 0.4793 (q25–q75 0.4498–0.5033)2,266nm — median 0.5111 (q25–q75 0.4807–0.5366)2,276nm — median 0.5404 (q25–q75 0.5105–0.5689)2,286nm — median 0.5572 (q25–q75 0.5281–0.5876)2,296nm — median 0.5573 (q25–q75 0.5293–0.5888)2,306nm — median 0.5666 (q25–q75 0.5397–0.5994)2,316nm — median 0.572 (q25–q75 0.5438–0.6039)2,328nm — median 0.5691 (q25–q75 0.5401–0.5997)2,338nm — median 0.565 (q25–q75 0.5363–0.596)2,348nm — median 0.5614 (q25–q75 0.5335–0.5927)2,358nm — median 0.556 (q25–q75 0.5272–0.5864)2,368nm — median 0.5547 (q25–q75 0.5257–0.585)2,378nm — median 0.5539 (q25–q75 0.5249–0.5842)2,388nm — median 0.5542 (q25–q75 0.5253–0.5845)2,398nm — median 0.5603 (q25–q75 0.5315–0.5911)2,408nm — median 0.5738 (q25–q75 0.5452–0.6052)2,418nm — median 0.591 (q25–q75 0.563–0.6234)2,428nm — median 0.6096 (q25–q75 0.5821–0.6425)2,438nm — median 0.6296 (q25–q75 0.6021–0.6619)2,448nm — median 0.6501 (q25–q75 0.6227–0.6822)2,458nm — median 0.6688 (q25–q75 0.641–0.7)2,468nm — median 0.6861 (q25–q75 0.6579–0.7169)2,478nm — median 0.6991 (q25–q75 0.6705–0.7292)2,488nm — median 0.705 (q25–q75 0.6762–0.7354)2,498nm — median 0.7044 (q25–q75 0.676–0.7347)

Sampling

Wavelengths700
Axis range1,100–2,498 nm
Mean spacing2 nm
Griduniform
Observations80

Signal & quality

Value range-0.0222 – 0.81
Mean range-0.00995 – 0.708
Mean level0.3449
Area482.2
PTP0.7175
Noise RMS7.4719e-05
SNR4.6e+03
SNR dB7e+01 dB
Dynamic range0.718
Smoothness0.0003019
Saturated0.0%
X-outliers27

Integrity & artefacts

NaN ratio0.00%
Inf count0
Zero ratio0.00%
Spike count1
Spike rate0.00%
Jump count0
Jump rate0.00%
Clip fraction0.00%

Shape & reference

Baseline slope0.60493
Curvature RMS0.00030202
D1 RMS0.0033234
RMS to mean0.022253
RMS p950.065094
SAM to mean0.011744
SAM p950.026485
Affine offset p950.019589
Affine gain p95 Δ0.11156
Affine residual p950.0042691
Xcorr lag p950

Outliers & repeatability

PCA Q p95/median9.8
Hotelling T2 p95/median8.6
Mahalanobis H p95/median2.9
Repeat groups0

Dimensionality (PCA)

Effective rank1.1
PCs → 95% var1
PCs → 99% var1
Top-10 cum. var100.0%
Computed metric scores 29worst 1.00
FamilleMétrique calculéeValeurScoreNiveauInterprétation datasetCauses typiquesCalcul / scoring
Intégrité des donnéesNaN ratiointegrity.nan_ratio0%0.00faibleSpectre completErreur acquisition/exportcount(isnan(X)) / X.sizealert = min(1, nan_ratio / 0.05)
Intégrité des donnéesInf countintegrity.inf_count00.00faibleNormalCalculs invalidescount(isinf(X))alert = min(1, inf_count / 1)
Intégrité des donnéesZero ratiointegrity.zero_ratio0%0.00faibleNormalExport, saturationcount(X == 0) / count(finite X)alert = min(1, zero_ratio / 0.05)
Amplitude globaleMean reflectanceamplitude.mean_reflectance0.344911.00fortValeur atypique: Trop clair / fond visible ou Trop sombreFond, géométriemean(X finite)alert reuses baseline/shape drift because absolute reflectance ranges are technology-dependent
Amplitude globaleArea under curveamplitude.area_under_curve482.181.00fortValeur atypique: Différence d'éclairement ou NormalDistance sondetrapezoid(mean_spectrum, spectral_axis)alert reuses baseline/shape drift because area scale depends on axis and units
Amplitude globalePeak-to-peak (PTP)amplitude.peak_to_peak0.71750.00faibleVariabilité forteSaturationmax(mean_spectrum) - min(mean_spectrum)alert increases when dynamic range is abnormally flat
Amplitude globaleVarianceamplitude.variance0.0360620.00faibleNormal ou hétérogèneMauvais contactvar(X finite)alert increases when variance/dynamic range is abnormally flat
BruitNoise RMSnoise.noise_rms7.4719e-050.00faibleStableLampe, détecteurmedian MAD(second derivative) * 1.4826 / sqrt(6)alert = noise_rms / signal_scale, saturated at 5%
BruitSNRnoise.snr4621.80.00faibleBon signalAcquisitionmean(abs(X)) / noise_rmsalert decreases with SNR dB; >=40 dB is treated as low alert
BruitBandwise SNRnoise.bandwise_snr_min13.9620.35faibleZone fiableDétecteurmin(abs(mean_spectrum) / local second-derivative noise)alert decreases with worst-band SNR dB; >=35 dB is treated as low alert
Artefacts locauxSpike countartefacts.spike_count10.00faibleSpectre propreCosmic rays, splicecount robust outliers in second derivativealert follows spike_rate, saturated at 1%
Artefacts locauxSpike rateartefacts.spike_rate0.00179%0.00faibleNormalInterpolationspike_count / (n_samples * (n_features - 2))alert = min(1, spike_rate / 0.01)
Artefacts locauxJump countartefacts.jump_count00.00faibleContinuSplicecount robust outliers in first derivativealert follows jump_rate, saturated at 1%
Artefacts locauxJump rateartefacts.jump_rate0%0.00faibleNormalCalibrationjump_count / (n_samples * (n_features - 1))alert = min(1, jump_rate / 0.01)
Artefacts locauxClip fractionartefacts.clip_fraction0.00357%0.00faibleNormalDétecteur saturéfraction of finite cells equal to repeated min/max extremaalert = min(1, clip_fraction / 0.01)
Forme spectraleBaseline slopeshape.baseline_slope0.604931.00fortDériveÉclairementlinear slope of mean_spectrum over normalized axisalert = abs(slope / signal_scale), saturated at 0.5
Forme spectraleCurvature RMSshape.curvature_rms0.000302020.04faibleLisseFond, splicemedian RMS(second derivative per spectrum)alert = curvature_rms / signal_scale, saturated at 1%
Forme spectraleD1 RMSshape.d1_rms0.00332340.09faiblePlatBiologie ou artefactmedian RMS(first derivative per spectrum)alert = d1_rms / signal_scale, saturated at 5%
Outliers multivariésPCA Q (SPE)outliers.pca_q_ratio9.76151.00fortSpectre atypiqueArtefact, mélangep95(Q/SPE residual) / median(Q/SPE residual)alert = min(1, pca_q_ratio / 8)
Outliers multivariésHotelling T²outliers.hotelling_t2_ratio8.61881.00fortExtrême mais cohérentVariabilité naturellep95(Hotelling T2) / median(Hotelling T2)alert = min(1, hotelling_t2_ratio / 8)
Outliers multivariésMahalanobis Houtliers.mahalanobis_h_ratio2.93810.73fortOutlier globalDomaine différentp95(sqrt(T2)) / median(sqrt(T2))alert = min(1, mahalanobis_h_ratio / 4)
Comparaison à référenceRMS to mean spectrumreference.rms_to_mean_spectrum_p950.0650940.36faibleTypiqueDomain shiftp95 RMS distance to dataset mean spectrumalert = RMS_p95 / signal_scale, saturated at 25%
Comparaison à référenceSpectral Angle Mapper (SAM)reference.sam_to_mean_spectrum_p950.0264850.08faibleSimilaireFond, géométriep95 spectral angle to dataset mean spectrumalert = min(1, SAM_p95 / 0.35 rad)
RépétabilitéRMS intra-IDrepeatability.rms_intra_id—0.00faibleStablePositionnementmedian RMS distance to repeated-sample centroidalert = RMS_intra_ID / signal_scale, saturated at 10%
RépétabilitéSAM intra-IDrepeatability.sam_intra_id—0.00faibleStableAcquisitionmedian SAM to repeated-sample centroidalert = min(1, SAM_intra_ID / 0.15 rad)
RépétabilitéCV intra-IDrepeatability.cv_intra_id—0.00faibleStableOpérateurmedian within-ID band CValert = min(1, CV_intra_ID / 0.25)
Structure du datasetPCA score densitystructure.pca_score_density12.0891.00fortSous-populationsLots différents1 / median kNN distance in PCA score spacealert follows density_cv/profile structure complexity, not raw density alone
Structure du datasetLocal Outlier Factor (LOF)structure.local_outlier_factor_p953.82831.00fortSpectre isoléCas raresp95 approximate LOF from PCA-score kNN distancesalert = min(1, max(0, LOF_p95 - 1) / 2)
Structure du datasetIsolation Forest scorestructure.isolation_forest_score_p950.614951.00fortSpectre atypiqueDiverses causesp95 IsolationForest anomaly score on PCA scoresalert follows structure complexity; raw score is implementation-dependent
X2 PCA score plot-2-10123-0.20.00.20.4PC1 -0.4696 · PC2 -0.005607PC1 -0.384 · PC2 -0.05502PC1 -0.802 · PC2 0.02677PC1 -0.9585 · PC2 -0.035PC1 0.4911 · PC2 0.01309PC1 0.7532 · PC2 -0.0528PC1 0.4598 · PC2 -0.01829PC1 0.6784 · PC2 -0.02081PC1 -0.2811 · PC2 -0.004903PC1 -0.4176 · PC2 -0.03321PC1 1.715 · PC2 0.02774PC1 1.513 · PC2 -0.0736PC1 -0.3795 · PC2 -0.05278PC1 -0.1555 · PC2 -0.001349PC1 1.502 · PC2 -0.03846PC1 1.861 · PC2 -0.1006PC1 0.3323 · PC2 0.03868PC1 0.6961 · PC2 0.001623PC1 0.4609 · PC2 0.05275PC1 0.9008 · PC2 -0.04073PC1 0.5271 · PC2 -0.05585PC1 0.9061 · PC2 -0.09007PC1 1.018 · PC2 -0.0809PC1 0.5551 · PC2 -0.03145PC1 0.6656 · PC2 0.08185PC1 0.1406 · PC2 0.006597PC1 1.844 · PC2 -0.1025PC1 1.429 · PC2 -0.1024PC1 1.225 · PC2 -0.05182PC1 -0.9599 · PC2 -0.03927PC1 -0.9765 · PC2 -0.04685PC1 -0.6111 · PC2 -0.0337PC1 0.1584 · PC2 -0.002515PC1 0.7507 · PC2 -0.101PC1 0.7124 · PC2 -0.01175PC1 -1.017 · PC2 0.0001711PC1 -1.113 · PC2 0.00713PC1 -0.995 · PC2 0.0149PC1 -0.8654 · PC2 -0.01645PC1 -1.408 · PC2 -0.04933PC1 -1.318 · PC2 -0.0115PC1 -1.32 · PC2 0.05107PC1 -1.362 · PC2 0.006507PC1 -0.5277 · PC2 0.002076PC1 -0.8078 · PC2 0.02144PC1 0.3731 · PC2 0.07147PC1 0.4574 · PC2 0.01779PC1 -0.3443 · PC2 0.07494PC1 -0.519 · PC2 0.03PC1 -0.5568 · PC2 0.02091PC1 0.01074 · PC2 0.02495PC1 0.04336 · PC2 0.04539PC1 0.278 · PC2 0.01893PC1 -1.488 · PC2 0.05126PC1 -1.524 · PC2 0.04677PC1 -0.7187 · PC2 -0.07126PC1 0.2427 · PC2 -0.1674PC1 0.1972 · PC2 -0.02015PC1 0.1604 · PC2 -0.0626PC1 0.1507 · PC2 -0.02525PC1 -0.6848 · PC2 0.009431PC1 0.1349 · PC2 0.047PC1 0.1901 · PC2 0.001298PC1 -0.5232 · PC2 -0.08039PC1 -0.3006 · PC2 -0.02507PC1 -1.03 · PC2 -0.03108PC1 -0.1956 · PC2 -0.06081PC1 -0.5609 · PC2 0.0008237PC1 -0.2786 · PC2 -0.04136PC1 -0.6168 · PC2 0.01382PC1 0.4159 · PC2 -0.1456PC1 -0.1227 · PC2 -0.1365PC1 -0.9394 · PC2 0.08018PC1 -1.152 · PC2 0.09742PC1 2.075 · PC2 0.2378PC1 -0.0673 · PC2 0.1628PC1 2.291 · PC2 0.2539PC1 0.006925 · PC2 0.1543PC1 0.1028 · PC2 0.2437PC1 0.3263 · PC2 0.1706PC1 (99.1%)PC2 (0.8%)80 scores
PCA explained variance0%25%50%75%100%PC1: 99.1% (cumulative 99.1%)1PC2: 0.8% (cumulative 99.8%)2PC3: 0.1% (cumulative 99.9%)3PC4: 0.0% (cumulative 99.9%)4PC5: 0.0% (cumulative 100.0%)5PC6: 0.0% (cumulative 100.0%)6PC7: 0.0% (cumulative 100.0%)7PC8: 0.0% (cumulative 100.0%)8PC9: 0.0% (cumulative 100.0%)9PC10: 0.0% (cumulative 100.0%)10cumulative explained variancePC variancecumulativeprincipal component · cumulative (dashed)
X-Y spectral correlation 4
X2 · Moisture spectral correlation-1-0.500.51absolute correlation envelopesigned correlationabsolute correlation1,0001,5002,0002,500|r|signed raxis · Pearson correlation scale
X2 · Oil spectral correlation-1-0.500.51absolute correlation envelopesigned correlationabsolute correlation1,0001,5002,0002,500|r|signed raxis · Pearson correlation scale
X2 · Protein spectral correlation-1-0.500.51absolute correlation envelopesigned correlationabsolute correlation1,0001,5002,0002,500|r|signed raxis · Pearson correlation scale
Targetmax |r|axis @ maxmean |r||r| ≥ .5
Moisture0.6662,4840.58896.3%
Oil0.3482,3080.2760.0%
Protein0.3592,3040.310.0%
Starch0.1791,1000.1190.0%

mp6spec

X3 · NIR · mp6
mp6spec spectra-0.250.000.250.500.751.001,0001,5002,0002,500q05-q95 envelopeq25-q75 envelopemedian spectrummedianq25–q75q05–q95wavelength / nm1,100nm — median -0.02038 (q25–q75 -0.02446–-0.01419)1,110nm — median -0.0205 (q25–q75 -0.02456–-0.01429)1,120nm — median -0.01894 (q25–q75 -0.02317–-0.01258)1,130nm — median -0.0132 (q25–q75 -0.01781–-0.005941)1,140nm — median -0.0003569 (q25–q75 -0.005702–0.008003)1,150nm — median 0.01756 (q25–q75 0.01129–0.02681)1,160nm — median 0.03589 (q25–q75 0.02749–0.0462)1,170nm — median 0.05053 (q25–q75 0.04151–0.06226)1,180nm — median 0.06311 (q25–q75 0.05347–0.07567)1,190nm — median 0.07447 (q25–q75 0.06398–0.0873)1,200nm — median 0.08068 (q25–q75 0.06982–0.09343)1,210nm — median 0.07954 (q25–q75 0.06876–0.09212)1,220nm — median 0.07169 (q25–q75 0.06136–0.08376)1,230nm — median 0.06182 (q25–q75 0.05232–0.07329)1,240nm — median 0.05436 (q25–q75 0.04558–0.06534)1,250nm — median 0.05015 (q25–q75 0.0417–0.06066)1,260nm — median 0.04796 (q25–q75 0.03965–0.05816)1,270nm — median 0.04629 (q25–q75 0.0381–0.05635)1,282nm — median 0.04364 (q25–q75 0.03567–0.05352)1,292nm — median 0.04135 (q25–q75 0.03356–0.05105)1,302nm — median 0.0406 (q25–q75 0.03282–0.05018)1,312nm — median 0.04256 (q25–q75 0.03462–0.05214)1,322nm — median 0.04823 (q25–q75 0.03974–0.058)1,332nm — median 0.05832 (q25–q75 0.04889–0.06857)1,342nm — median 0.07239 (q25–q75 0.06185–0.08332)1,352nm — median 0.08849 (q25–q75 0.07707–0.1001)1,362nm — median 0.1038 (q25–q75 0.09123–0.1161)1,372nm — median 0.1165 (q25–q75 0.1029–0.1292)1,382nm — median 0.1297 (q25–q75 0.1149–0.1431)1,392nm — median 0.1506 (q25–q75 0.1341–0.1656)1,402nm — median 0.1837 (q25–q75 0.1648–0.2)1,412nm — median 0.2283 (q25–q75 0.2073–0.2455)1,422nm — median 0.2749 (q25–q75 0.2516–0.2929)1,432nm — median 0.3118 (q25–q75 0.2874–0.3314)1,442nm — median 0.3335 (q25–q75 0.3083–0.3539)1,452nm — median 0.3433 (q25–q75 0.3178–0.3639)1,462nm — median 0.3473 (q25–q75 0.3216–0.3678)1,472nm — median 0.3473 (q25–q75 0.3217–0.3676)1,482nm — median 0.3438 (q25–q75 0.3185–0.364)1,492nm — median 0.3385 (q25–q75 0.3136–0.3586)1,502nm — median 0.3318 (q25–q75 0.3073–0.352)1,512nm — median 0.3248 (q25–q75 0.3002–0.3446)1,522nm — median 0.3184 (q25–q75 0.2939–0.3377)1,532nm — median 0.3135 (q25–q75 0.289–0.3323)1,542nm — median 0.3105 (q25–q75 0.2859–0.3288)1,552nm — median 0.3085 (q25–q75 0.2839–0.3265)1,562nm — median 0.307 (q25–q75 0.2823–0.3248)1,572nm — median 0.3047 (q25–q75 0.2801–0.3224)1,582nm — median 0.3002 (q25–q75 0.2759–0.318)1,592nm — median 0.2929 (q25–q75 0.2691–0.3106)1,602nm — median 0.2831 (q25–q75 0.2597–0.3006)1,612nm — median 0.2724 (q25–q75 0.2493–0.2895)1,622nm — median 0.2617 (q25–q75 0.2393–0.2786)1,634nm — median 0.25 (q25–q75 0.228–0.2666)1,644nm — median 0.2415 (q25–q75 0.22–0.258)1,654nm — median 0.235 (q25–q75 0.2139–0.2514)1,664nm — median 0.2314 (q25–q75 0.2108–0.2481)1,674nm — median 0.2321 (q25–q75 0.2118–0.2492)1,684nm — median 0.2411 (q25–q75 0.2204–0.2591)1,694nm — median 0.2572 (q25–q75 0.236–0.276)1,704nm — median 0.2687 (q25–q75 0.2475–0.288)1,714nm — median 0.2724 (q25–q75 0.2512–0.2921)1,724nm — median 0.2746 (q25–q75 0.2532–0.2946)1,734nm — median 0.2755 (q25–q75 0.2537–0.2951)1,744nm — median 0.2781 (q25–q75 0.2558–0.2973)1,754nm — median 0.282 (q25–q75 0.2597–0.3012)1,764nm — median 0.2836 (q25–q75 0.2612–0.3027)1,774nm — median 0.2838 (q25–q75 0.261–0.3024)1,784nm — median 0.2808 (q25–q75 0.258–0.2993)1,794nm — median 0.2736 (q25–q75 0.2513–0.2922)1,804nm — median 0.2654 (q25–q75 0.2435–0.2841)1,814nm — median 0.259 (q25–q75 0.2374–0.2777)1,824nm — median 0.2535 (q25–q75 0.2321–0.2721)1,834nm — median 0.2466 (q25–q75 0.2256–0.265)1,844nm — median 0.239 (q25–q75 0.2184–0.2572)1,854nm — median 0.2353 (q25–q75 0.2148–0.2536)1,864nm — median 0.2393 (q25–q75 0.2184–0.2582)1,874nm — median 0.2542 (q25–q75 0.2321–0.2734)1,884nm — median 0.2861 (q25–q75 0.2622–0.305)1,894nm — median 0.3411 (q25–q75 0.3166–0.3636)1,904nm — median 0.4139 (q25–q75 0.3873–0.4389)1,914nm — median 0.4767 (q25–q75 0.4481–0.504)1,924nm — median 0.5156 (q25–q75 0.4865–0.5447)1,934nm — median 0.5284 (q25–q75 0.4986–0.5577)1,944nm — median 0.5165 (q25–q75 0.4871–0.5459)1,954nm — median 0.495 (q25–q75 0.4665–0.5242)1,964nm — median 0.4737 (q25–q75 0.4457–0.5021)1,976nm — median 0.4499 (q25–q75 0.4218–0.4774)1,986nm — median 0.4306 (q25–q75 0.4027–0.4575)1,996nm — median 0.4145 (q25–q75 0.3872–0.4408)2,006nm — median 0.4056 (q25–q75 0.3786–0.4312)2,016nm — median 0.4062 (q25–q75 0.3794–0.4314)2,026nm — median 0.417 (q25–q75 0.3901–0.4421)2,036nm — median 0.4356 (q25–q75 0.4086–0.4607)2,046nm — median 0.4574 (q25–q75 0.4303–0.483)2,056nm — median 0.4782 (q25–q75 0.4505–0.5039)2,066nm — median 0.4951 (q25–q75 0.4667–0.5199)2,076nm — median 0.508 (q25–q75 0.4792–0.5318)2,086nm — median 0.5175 (q25–q75 0.4881–0.5411)2,096nm — median 0.5235 (q25–q75 0.4937–0.5471)2,106nm — median 0.5258 (q25–q75 0.4956–0.5494)2,116nm — median 0.5241 (q25–q75 0.4937–0.5476)2,126nm — median 0.5181 (q25–q75 0.4877–0.5413)2,136nm — median 0.5084 (q25–q75 0.4781–0.5315)2,146nm — median 0.498 (q25–q75 0.4678–0.5212)2,156nm — median 0.4861 (q25–q75 0.456–0.5099)2,166nm — median 0.4739 (q25–q75 0.444–0.4984)2,176nm — median 0.4613 (q25–q75 0.432–0.4864)2,186nm — median 0.4489 (q25–q75 0.42–0.4739)2,196nm — median 0.4378 (q25–q75 0.4091–0.4622)2,206nm — median 0.4277 (q25–q75 0.3993–0.4516)2,216nm — median 0.419 (q25–q75 0.3912–0.4428)2,226nm — median 0.4145 (q25–q75 0.3874–0.4384)2,236nm — median 0.4188 (q25–q75 0.3922–0.4435)2,246nm — median 0.4358 (q25–q75 0.4089–0.4619)2,256nm — median 0.4646 (q25–q75 0.4366–0.4918)2,266nm — median 0.497 (q25–q75 0.4674–0.5245)2,276nm — median 0.5275 (q25–q75 0.4981–0.5563)2,286nm — median 0.5462 (q25–q75 0.5167–0.5757)2,296nm — median 0.5458 (q25–q75 0.5167–0.5776)2,306nm — median 0.5556 (q25–q75 0.5266–0.5888)2,316nm — median 0.5608 (q25–q75 0.5316–0.5924)2,328nm — median 0.5576 (q25–q75 0.5283–0.5884)2,338nm — median 0.553 (q25–q75 0.5241–0.585)2,348nm — median 0.5491 (q25–q75 0.5204–0.5824)2,358nm — median 0.5431 (q25–q75 0.5141–0.5751)2,368nm — median 0.5424 (q25–q75 0.5134–0.5741)2,378nm — median 0.5415 (q25–q75 0.5127–0.5733)2,388nm — median 0.5414 (q25–q75 0.5126–0.5732)2,398nm — median 0.5477 (q25–q75 0.5188–0.5793)2,408nm — median 0.5624 (q25–q75 0.533–0.5934)2,418nm — median 0.5806 (q25–q75 0.5508–0.6109)2,428nm — median 0.5991 (q25–q75 0.5691–0.6289)2,438nm — median 0.6194 (q25–q75 0.5892–0.6488)2,448nm — median 0.6409 (q25–q75 0.6103–0.6698)2,458nm — median 0.6597 (q25–q75 0.6293–0.6884)2,468nm — median 0.6761 (q25–q75 0.6462–0.7048)2,478nm — median 0.6889 (q25–q75 0.659–0.7177)2,488nm — median 0.6952 (q25–q75 0.6653–0.7239)2,498nm — median 0.6945 (q25–q75 0.6645–0.723)

Sampling

Wavelengths700
Axis range1,100–2,498 nm
Mean spacing2 nm
Griduniform
Observations80

Signal & quality

Value range-0.0339 – 0.801
Mean range-0.0195 – 0.697
Mean level0.3332
Area465.7
PTP0.7169
Noise RMS7.5667e-05
SNR4.4e+03
SNR dB7e+01 dB
Dynamic range0.717
Smoothness0.0003049
Saturated0.0%
X-outliers29

Integrity & artefacts

NaN ratio0.00%
Inf count0
Zero ratio0.00%
Spike count2
Spike rate0.00%
Jump count0
Jump rate0.00%
Clip fraction0.00%

Shape & reference

Baseline slope0.6017
Curvature RMS0.00030523
D1 RMS0.0033757
RMS to mean0.02345
RMS p950.064206
SAM to mean0.013967
SAM p950.031301
Affine offset p950.020558
Affine gain p95 Δ0.11484
Affine residual p950.004491
Xcorr lag p950

Outliers & repeatability

PCA Q p95/median11
Hotelling T2 p95/median7.5
Mahalanobis H p95/median2.7
Repeat groups0

Dimensionality (PCA)

Effective rank1.1
PCs → 95% var1
PCs → 99% var1
Top-10 cum. var100.0%
Computed metric scores 29worst 1.00
FamilleMétrique calculéeValeurScoreNiveauInterprétation datasetCauses typiquesCalcul / scoring
Intégrité des donnéesNaN ratiointegrity.nan_ratio0%0.00faibleSpectre completErreur acquisition/exportcount(isnan(X)) / X.sizealert = min(1, nan_ratio / 0.05)
Intégrité des donnéesInf countintegrity.inf_count00.00faibleNormalCalculs invalidescount(isinf(X))alert = min(1, inf_count / 1)
Intégrité des donnéesZero ratiointegrity.zero_ratio0%0.00faibleNormalExport, saturationcount(X == 0) / count(finite X)alert = min(1, zero_ratio / 0.05)
Amplitude globaleMean reflectanceamplitude.mean_reflectance0.333151.00fortValeur atypique: Trop clair / fond visible ou Trop sombreFond, géométriemean(X finite)alert reuses baseline/shape drift because absolute reflectance ranges are technology-dependent
Amplitude globaleArea under curveamplitude.area_under_curve465.741.00fortValeur atypique: Différence d'éclairement ou NormalDistance sondetrapezoid(mean_spectrum, spectral_axis)alert reuses baseline/shape drift because area scale depends on axis and units
Amplitude globalePeak-to-peak (PTP)amplitude.peak_to_peak0.716940.00faibleVariabilité forteSaturationmax(mean_spectrum) - min(mean_spectrum)alert increases when dynamic range is abnormally flat
Amplitude globaleVarianceamplitude.variance0.0359240.00faibleNormal ou hétérogèneMauvais contactvar(X finite)alert increases when variance/dynamic range is abnormally flat
BruitNoise RMSnoise.noise_rms7.5667e-050.00faibleStableLampe, détecteurmedian MAD(second derivative) * 1.4826 / sqrt(6)alert = noise_rms / signal_scale, saturated at 5%
BruitSNRnoise.snr44150.00faibleBon signalAcquisitionmean(abs(X)) / noise_rmsalert decreases with SNR dB; >=40 dB is treated as low alert
BruitBandwise SNRnoise.bandwise_snr_min113.730.00faibleZone fiableDétecteurmin(abs(mean_spectrum) / local second-derivative noise)alert decreases with worst-band SNR dB; >=35 dB is treated as low alert
Artefacts locauxSpike countartefacts.spike_count20.00faibleSpectre propreCosmic rays, splicecount robust outliers in second derivativealert follows spike_rate, saturated at 1%
Artefacts locauxSpike rateartefacts.spike_rate0.00358%0.00faibleNormalInterpolationspike_count / (n_samples * (n_features - 2))alert = min(1, spike_rate / 0.01)
Artefacts locauxJump countartefacts.jump_count00.00faibleContinuSplicecount robust outliers in first derivativealert follows jump_rate, saturated at 1%
Artefacts locauxJump rateartefacts.jump_rate0%0.00faibleNormalCalibrationjump_count / (n_samples * (n_features - 1))alert = min(1, jump_rate / 0.01)
Artefacts locauxClip fractionartefacts.clip_fraction0.00357%0.00faibleNormalDétecteur saturéfraction of finite cells equal to repeated min/max extremaalert = min(1, clip_fraction / 0.01)
Forme spectraleBaseline slopeshape.baseline_slope0.60171.00fortDériveÉclairementlinear slope of mean_spectrum over normalized axisalert = abs(slope / signal_scale), saturated at 0.5
Forme spectraleCurvature RMSshape.curvature_rms0.000305230.04faibleLisseFond, splicemedian RMS(second derivative per spectrum)alert = curvature_rms / signal_scale, saturated at 1%
Forme spectraleD1 RMSshape.d1_rms0.00337570.09faiblePlatBiologie ou artefactmedian RMS(first derivative per spectrum)alert = d1_rms / signal_scale, saturated at 5%
Outliers multivariésPCA Q (SPE)outliers.pca_q_ratio11.4161.00fortSpectre atypiqueArtefact, mélangep95(Q/SPE residual) / median(Q/SPE residual)alert = min(1, pca_q_ratio / 8)
Outliers multivariésHotelling T²outliers.hotelling_t2_ratio7.53580.94fortExtrême mais cohérentVariabilité naturellep95(Hotelling T2) / median(Hotelling T2)alert = min(1, hotelling_t2_ratio / 8)
Outliers multivariésMahalanobis Houtliers.mahalanobis_h_ratio2.74430.69moyenOutlier globalDomaine différentp95(sqrt(T2)) / median(sqrt(T2))alert = min(1, mahalanobis_h_ratio / 4)
Comparaison à référenceRMS to mean spectrumreference.rms_to_mean_spectrum_p950.0642060.36faibleTypiqueDomain shiftp95 RMS distance to dataset mean spectrumalert = RMS_p95 / signal_scale, saturated at 25%
Comparaison à référenceSpectral Angle Mapper (SAM)reference.sam_to_mean_spectrum_p950.0313010.09faibleSimilaireFond, géométriep95 spectral angle to dataset mean spectrumalert = min(1, SAM_p95 / 0.35 rad)
RépétabilitéRMS intra-IDrepeatability.rms_intra_id—0.00faibleStablePositionnementmedian RMS distance to repeated-sample centroidalert = RMS_intra_ID / signal_scale, saturated at 10%
RépétabilitéSAM intra-IDrepeatability.sam_intra_id—0.00faibleStableAcquisitionmedian SAM to repeated-sample centroidalert = min(1, SAM_intra_ID / 0.15 rad)
RépétabilitéCV intra-IDrepeatability.cv_intra_id—0.00faibleStableOpérateurmedian within-ID band CValert = min(1, CV_intra_ID / 0.25)
Structure du datasetPCA score densitystructure.pca_score_density11.7841.00fortSous-populationsLots différents1 / median kNN distance in PCA score spacealert follows density_cv/profile structure complexity, not raw density alone
Structure du datasetLocal Outlier Factor (LOF)structure.local_outlier_factor_p953.67481.00fortSpectre isoléCas raresp95 approximate LOF from PCA-score kNN distancesalert = min(1, max(0, LOF_p95 - 1) / 2)
Structure du datasetIsolation Forest scorestructure.isolation_forest_score_p950.622931.00fortSpectre atypiqueDiverses causesp95 IsolationForest anomaly score on PCA scoresalert follows structure complexity; raw score is implementation-dependent
X3 PCA score plot-2-10123-0.4-0.20.00.20.4PC1 -0.4846 · PC2 -0.0157PC1 -0.3744 · PC2 -0.03885PC1 -0.8288 · PC2 0.04457PC1 -0.9792 · PC2 -0.004972PC1 0.5153 · PC2 0.03459PC1 0.6619 · PC2 -0.03222PC1 0.4784 · PC2 0.008005PC1 0.68 · PC2 -0.009673PC1 -0.3275 · PC2 0.02315PC1 -0.4073 · PC2 -0.009215PC1 1.688 · PC2 0.02587PC1 1.567 · PC2 -0.06987PC1 -0.3919 · PC2 -0.0287PC1 -0.1232 · PC2 0.01996PC1 1.569 · PC2 -0.02912PC1 1.909 · PC2 -0.08584PC1 0.3192 · PC2 0.05962PC1 0.7143 · PC2 0.02043PC1 0.4811 · PC2 0.07054PC1 0.9316 · PC2 -0.02025PC1 0.5107 · PC2 -0.02833PC1 0.9631 · PC2 -0.06758PC1 1.039 · PC2 -0.0685PC1 0.55 · PC2 -0.02031PC1 0.6523 · PC2 0.08795PC1 0.1546 · PC2 0.02195PC1 1.888 · PC2 -0.08746PC1 1.508 · PC2 -0.09037PC1 1.262 · PC2 -0.02999PC1 -0.986 · PC2 -0.001168PC1 -0.9718 · PC2 -0.007064PC1 -0.5904 · PC2 -0.007741PC1 0.1658 · PC2 0.01634PC1 0.9069 · PC2 -0.04569PC1 0.8162 · PC2 0.01953PC1 -0.9855 · PC2 0.03126PC1 -1.075 · PC2 0.04304PC1 -0.9201 · PC2 0.05082PC1 -0.8244 · PC2 0.01071PC1 -1.365 · PC2 -0.01796PC1 -1.356 · PC2 -0.01507PC1 -1.344 · PC2 0.03102PC1 -1.396 · PC2 0.002128PC1 -0.5352 · PC2 -0.02188PC1 -0.8074 · PC2 -0.006933PC1 0.3022 · PC2 0.04316PC1 0.4284 · PC2 -0.02342PC1 -0.3874 · PC2 0.05793PC1 -0.5582 · PC2 0.006785PC1 -0.6166 · PC2 0.002398PC1 -0.03282 · PC2 0.005359PC1 -0.0268 · PC2 0.01806PC1 0.1913 · PC2 -0.001443PC1 -1.56 · PC2 0.03755PC1 -1.543 · PC2 0.043PC1 -0.8428 · PC2 -0.1199PC1 0.2122 · PC2 -0.2006PC1 0.06604 · PC2 -0.03077PC1 0.1187 · PC2 -0.0814PC1 0.1474 · PC2 -0.0443PC1 -0.6211 · PC2 -0.02386PC1 0.1031 · PC2 0.01556PC1 0.1473 · PC2 -0.01946PC1 -0.5419 · PC2 -0.1002PC1 -0.3368 · PC2 -0.04598PC1 -1.042 · PC2 -0.05914PC1 -0.2249 · PC2 -0.08744PC1 -0.5952 · PC2 -0.01931PC1 -0.2653 · PC2 -0.05975PC1 -0.6675 · PC2 0.007617PC1 0.6791 · PC2 -0.1281PC1 -0.13 · PC2 -0.1687PC1 -0.9396 · PC2 0.06122PC1 -1.215 · PC2 0.1014PC1 2.215 · PC2 0.2145PC1 -0.1525 · PC2 0.1524PC1 2.255 · PC2 0.1901PC1 0.1343 · PC2 0.127PC1 0.09911 · PC2 0.2087PC1 0.3412 · PC2 0.1601PC1 (99.2%)PC2 (0.6%)80 scores
PCA explained variance0%25%50%75%100%PC1: 99.2% (cumulative 99.2%)1PC2: 0.6% (cumulative 99.8%)2PC3: 0.1% (cumulative 99.9%)3PC4: 0.0% (cumulative 99.9%)4PC5: 0.0% (cumulative 100.0%)5PC6: 0.0% (cumulative 100.0%)6PC7: 0.0% (cumulative 100.0%)7PC8: 0.0% (cumulative 100.0%)8PC9: 0.0% (cumulative 100.0%)9PC10: 0.0% (cumulative 100.0%)10cumulative explained variancePC variancecumulativeprincipal component · cumulative (dashed)
X-Y spectral correlation 4
X3 · Moisture spectral correlation-1-0.500.51absolute correlation envelopesigned correlationabsolute correlation1,0001,5002,0002,500|r|signed raxis · Pearson correlation scale
X3 · Oil spectral correlation-1-0.500.51absolute correlation envelopesigned correlationabsolute correlation1,0001,5002,0002,500|r|signed raxis · Pearson correlation scale
X3 · Protein spectral correlation-1-0.500.51absolute correlation envelopesigned correlationabsolute correlation1,0001,5002,0002,500|r|signed raxis · Pearson correlation scale
Targetmax |r|axis @ maxmean |r||r| ≥ .5
Moisture0.6672,4780.59297.6%
Oil0.3442,3060.2880.0%
Protein0.3572,3040.3240.0%
Starch0.1271,4440.1010.0%

Metric interpretation reference

Metric catalog 29
FamilleMétriqueCe qu’elle détecteForte valeur =Faible valeur =Causes typiquesCalcul / score
Intégrité des donnéesNaN ratioDonnées manquantesSpectre corrompuSpectre completErreur acquisition/exportcount(isnan(X)) / X.sizealert = min(1, nan_ratio / 0.05)
Intégrité des donnéesInf countValeurs infiniesCorruptionNormalCalculs invalidescount(isinf(X))alert = min(1, inf_count / 1)
Intégrité des donnéesZero ratioColonnes ou cellules nullesSpectre tronquéNormalExport, saturationcount(X == 0) / count(finite X)alert = min(1, zero_ratio / 0.05)
Amplitude globaleMean reflectanceNiveau moyenTrop clair / fond visibleTrop sombreFond, géométriemean(X finite)alert reuses baseline/shape drift because absolute reflectance ranges are technology-dependent
Amplitude globaleArea under curveIntensité globaleDifférence d'éclairementNormalDistance sondetrapezoid(mean_spectrum, spectral_axis)alert reuses baseline/shape drift because area scale depends on axis and units
Amplitude globalePeak-to-peak (PTP)DynamiqueVariabilité forteSpectre platSaturationmax(mean_spectrum) - min(mean_spectrum)alert increases when dynamic range is abnormally flat
Amplitude globaleVarianceVariabilité spectraleNormal ou hétérogèneSpectre platMauvais contactvar(X finite)alert increases when variance/dynamic range is abnormally flat
BruitNoise RMSBruit haute fréquenceBruitéStableLampe, détecteurmedian MAD(second derivative) * 1.4826 / sqrt(6)alert = noise_rms / signal_scale, saturated at 5%
BruitSNRQualité signalBon signalMauvais signalAcquisitionmean(abs(X)) / noise_rmsalert decreases with SNR dB; >=40 dB is treated as low alert
BruitBandwise SNRBruit localiséZone fiableZone problématiqueDétecteurmin(abs(mean_spectrum) / local second-derivative noise)alert decreases with worst-band SNR dB; >=35 dB is treated as low alert
Artefacts locauxSpike countPics étroitsArtefactsSpectre propreCosmic rays, splicecount robust outliers in second derivativealert follows spike_rate, saturated at 1%
Artefacts locauxSpike rateDensité de picsSpectre suspectNormalInterpolationspike_count / (n_samples * (n_features - 2))alert = min(1, spike_rate / 0.01)
Artefacts locauxJump countDiscontinuitésRaccord détecteurContinuSplicecount robust outliers in first derivativealert follows jump_rate, saturated at 1%
Artefacts locauxJump rateFréquence de sautsProblème spectralNormalCalibrationjump_count / (n_samples * (n_features - 1))alert = min(1, jump_rate / 0.01)
Artefacts locauxClip fractionSaturationClippingNormalDétecteur saturéfraction of finite cells equal to repeated min/max extremaalert = min(1, clip_fraction / 0.01)
Forme spectraleBaseline slopePente globaleDériveStableÉclairementlinear slope of mean_spectrum over normalized axisalert = abs(slope / signal_scale), saturated at 0.5
Forme spectraleCurvature RMSCourbureForme inhabituelleLisseFond, splicemedian RMS(second derivative per spectrum)alert = curvature_rms / signal_scale, saturated at 1%
Forme spectraleD1 RMSVariabilité localeSpectre structuréPlatBiologie ou artefactmedian RMS(first derivative per spectrum)alert = d1_rms / signal_scale, saturated at 5%
Outliers multivariésPCA Q (SPE)Non expliqué par PCASpectre atypiqueConformeArtefact, mélangep95(Q/SPE residual) / median(Q/SPE residual)alert = min(1, pca_q_ratio / 8)
Outliers multivariésHotelling T²Extrême dans PCAExtrême mais cohérentCentralVariabilité naturellep95(Hotelling T2) / median(Hotelling T2)alert = min(1, hotelling_t2_ratio / 8)
Outliers multivariésMahalanobis HDistance au nuageOutlier globalPopulation normaleDomaine différentp95(sqrt(T2)) / median(sqrt(T2))alert = min(1, mahalanobis_h_ratio / 4)
Comparaison à référenceRMS to mean spectrumDistance moyenneSpectre différentTypiqueDomain shiftp95 RMS distance to dataset mean spectrumalert = RMS_p95 / signal_scale, saturated at 25%
Comparaison à référenceSpectral Angle Mapper (SAM)Différence de formeForme différenteSimilaireFond, géométriep95 spectral angle to dataset mean spectrumalert = min(1, SAM_p95 / 0.35 rad)
RépétabilitéRMS intra-IDReproductibilitéMauvaise répétabilitéStablePositionnementmedian RMS distance to repeated-sample centroidalert = RMS_intra_ID / signal_scale, saturated at 10%
RépétabilitéSAM intra-IDVariation de formeInstableStableAcquisitionmedian SAM to repeated-sample centroidalert = min(1, SAM_intra_ID / 0.15 rad)
RépétabilitéCV intra-IDVariabilité interneMauvais contrôleStableOpérateurmedian within-ID band CValert = min(1, CV_intra_ID / 0.25)
Structure du datasetPCA score densityClustersSous-populationsHomogèneLots différents1 / median kNN distance in PCA score spacealert follows density_cv/profile structure complexity, not raw density alone
Structure du datasetLocal Outlier Factor (LOF)Anomalie localeSpectre isoléPopulation normaleCas raresp95 approximate LOF from PCA-score kNN distancesalert = min(1, max(0, LOF_p95 - 1) / 2)
Structure du datasetIsolation Forest scoreAnomalie globaleSpectre atypiqueNormalDiverses causesp95 IsolationForest anomaly score on PCA scoresalert follows structure complexity; raw score is implementation-dependent
Technology-specific extensions
TechnologieAdaptations / métriquesAnomalies cibléesCommentaire pratique
UV-Vis 300-1000 nmBaseline, pente globale, dérive aux bords 300-350 et 900-1000; métriques par zonesLumière parasite, mauvais blanc, saturation, faible signal aux extrémitésLes bords sont souvent instables; calculer aussi des scores edge/middle.
UV-Vis 300-1000 nmSaturation / clipping proche absorbance max ou réflectance maxSignal écrêtéTrès important si absorption forte.
UV-Vis 300-1000 nmRed-edge, position de maximum, ratios de bandes si végétalDécalage biologique ou artefact optiqueAide à distinguer changement réel et problème d'acquisition.
UV-Vis 300-1000 nmSmoothness / roughness indexBruit haute fréquenceSouvent plus informatif que le SNR seul.
MIR / ATR-FTIRATR contact quality index: intensité globale, aire totale, profondeur des bandes clésMauvais contact cristal-échantillonCrucial: beaucoup d'anomalies viennent du contact ATR.
MIR / ATR-FTIRCO2 / H2O atmospheric bandsMauvaise correction atmosphériquePics parasites fréquents.
MIR / ATR-FTIRBaseline curvature / rubber-band residualDiffusion, contact, dérive baselineTrès utile avant PCA.
MIR / ATR-FTIRPeak position shiftMauvais alignement spectral / calibrationImportant en FTIR car de petits shifts comptent.
MIR / ATR-FTIRBand area ratios sur bandes connuesSpectre chimiquement incohérentÀ adapter par matrice: polysaccharides, protéines, lipides, etc.
HS-MSTotal Ion Current (TIC), Base Peak Intensity (BPI)Injection faible, ionisation instableÉquivalent MS du niveau global spectral.
HS-MSNombre de pics détectésSpectre pauvre ou trop bruitéTrop peu = mauvais signal; trop = bruit/contamination.
HS-MSMass accuracy / m/z driftProblème calibration masseFondamental en HRMS.
HS-MSRetention time drift si LC/GC-MSDérive chromatographiqueÀ suivre sur standards/QC pools.
HS-MSBlank contamination scoreContaminants / carry-overComparer échantillons vs blancs.
HS-MSInternal standard CVVariabilité instrumentaleTrès robuste si standards disponibles.
HS-MSMissingness par featureInstabilité de détectionCrucial pour filtrer les variables.
Avec répétitionsRMS intra-échantillonRépétabilité globaleApplicable à toutes les technologies.
Avec répétitionsSAM / corrélation intra-échantillonRépétabilité de formeTrès utile pour spectres.
Avec répétitionsCV intra-échantillon par bande / featureRépétabilité localeDétecte les zones instables.
Avec répétitionsICC ou variance componentsPart variance échantillon vs techniqueTrès utile si plusieurs répétitions par sample.
Avec répétitionsDistance au centroïde intra-IDRépétition aberrantePermet de flagger la mauvaise répétition plutôt que le sample entier.
Bug-hunting / supervised audits
Famille de bug potentielMéthodes à ajouterCe que ça détecteÉtat dans l’explorateur
Shift spectral globalCorrélation spectre moyen inter-dataset, DTW, cross-correlation, comparaison positions de picsDécalage en longueur d'onde, mauvais alignement, interpolation différentePartiellement calculé: cross-correlation lag et dispersion des positions de pics vs spectre moyen.
Baseline / offset / gainRégression chaque spectre vs spectre moyen: x = a + b ref + residual; suivi de a, b, RMS résiduelOffset additif, effet multiplicatif, dérive de baselineCalculé dans reference.affine_*.
Mélange de lignes / mauvais appariement X-M-YVérification index, hash des lignes, duplication ID, distance spectrale intra-ID, labels incohérentsLignes mélangées, metadata mal alignées, Y attribué au mauvais spectrePartiellement couvert par répétabilité intra-ID; checks index/hash à ajouter au pipeline canonical.
Fuite d'information / répétitions mal splitéesGroupKFold par sample_id vs StratifiedKFold random; audit des partitions par sample_idPerformance artificiellement bonne due aux répétitionsNécessite splits et benchmark modèle; non calculé par la carte descriptive.
Label bugsÉchantillons proches en X mais Y différents, confident learning, erreurs systématiques FP/FNY inversés, erreurs de saisie, classes ambiguësNécessite Y et/ou modèle; recommandé pour l'explorateur supervisé.
Sous-domaines cachésPCA/UMAP/t-SNE + clustering non supervisé + association avec dataset/Y/date/operatorLots, campagnes, sondes, backgrounds non renseignésPartiellement calculé par structure PCA/LOF; UMAP/t-SNE hors carte statique.
Artefacts localisés inconnusCarte wavelength x dataset: différence moyenne, différence variance, KS par longueur d'ondeRégions spectrales anormales non anticipéesÀ calculer au niveau banque quand plusieurs datasets partagent un axe spectral.
Ruptures instrumentalesDiscontinuités dans dérivées, changepoint detectionSplice, raccord détecteur, saut local non prévuCalculé par jump/spike rates; changepoint plus avancé à ajouter.
Mélange / contamination spectraleNMF / unmixing / reconstruction par convex hullComposante externe: fond, plastique, solNon calculé automatiquement; nécessite hypothèses de composants ou grande bibliothèque.
Features instables mais prédictivesImportance modèle vs instabilité QC par variableModèle qui apprend un artefact plutôt qu'un signal biologiqueNécessite modèle supervisé; recommandé pour rapports de benchmark.

Variables

Targets 4

Moisture

target · numeric
Moisture distribution0510159.377 – 9.444: 39.444 – 9.512: 09.512 – 9.579: 09.579 – 9.646: 19.646 – 9.714: 29.714 – 9.781: 69.781 – 9.848: 29.848 – 9.916: 39.916 – 9.983: 69.983 – 10.05: 210.05 – 10.12: 410.12 – 10.19: 710.19 – 10.25: 110.25 – 10.32: 1110.32 – 10.39: 610.39 – 10.45: 610.45 – 10.52: 110.52 – 10.59: 510.59 – 10.66: 310.66 – 10.72: 110.72 – 10.79: 210.79 – 10.86: 410.86 – 10.93: 110.93 – 10.99: 39.09.510.010.511.0
n / missing80 / 0
Mean ± SD10.23 ± 0.38
Median10.28
Range9.377 – 10.99
CV0.0372
Skew / kurtosis-0.05 / -0.43
Normal?yes

Oil

target · numeric
Oil distribution05103.088 – 3.119: 23.119 – 3.15: 03.15 – 3.181: 23.181 – 3.212: 13.212 – 3.243: 33.243 – 3.274: 23.274 – 3.305: 23.305 – 3.336: 43.336 – 3.367: 23.367 – 3.398: 13.398 – 3.429: 53.429 – 3.46: 83.46 – 3.491: 43.491 – 3.522: 93.522 – 3.553: 43.553 – 3.584: 53.584 – 3.615: 43.615 – 3.646: 33.646 – 3.677: 43.677 – 3.708: 53.708 – 3.739: 33.739 – 3.77: 33.77 – 3.801: 13.801 – 3.832: 33.03.23.43.63.84.0
n / missing80 / 0
Mean ± SD3.498 ± 0.177
Median3.497
Range3.088 – 3.832
CV0.0506
Skew / kurtosis-0.24 / -0.46
Normal?yes

Protein

target · numeric
Protein distribution05107.654 – 7.74: 17.74 – 7.825: 27.825 – 7.911: 37.911 – 7.997: 07.997 – 8.083: 18.083 – 8.168: 48.168 – 8.254: 28.254 – 8.34: 98.34 – 8.425: 68.425 – 8.511: 108.511 – 8.597: 48.597 – 8.683: 38.683 – 8.768: 68.768 – 8.854: 38.854 – 8.94: 28.94 – 9.025: 59.025 – 9.111: 29.111 – 9.197: 29.197 – 9.282: 09.282 – 9.368: 59.368 – 9.454: 69.454 – 9.54: 09.54 – 9.625: 29.625 – 9.711: 278910
n / missing80 / 0
Mean ± SD8.668 ± 0.499
Median8.561
Range7.654 – 9.711
CV0.0575
Skew / kurtosis0.28 / -0.65
Normal?yes

Starch

target · numeric
Starch distribution051062.83 – 62.98: 262.98 – 63.13: 263.13 – 63.28: 363.28 – 63.43: 263.43 – 63.59: 363.59 – 63.74: 163.74 – 63.89: 263.89 – 64.04: 164.04 – 64.19: 264.19 – 64.35: 564.35 – 64.5: 364.5 – 64.65: 964.65 – 64.8: 464.8 – 64.95: 964.95 – 65.1: 265.1 – 65.26: 865.26 – 65.41: 565.41 – 65.56: 765.56 – 65.71: 465.71 – 65.86: 465.86 – 66.02: 166.02 – 66.17: 066.17 – 66.32: 066.32 – 66.47: 1626364656667
n / missing80 / 0
Mean ± SD64.7 ± 0.821
Median64.82
Range62.83 – 66.47
CV0.0127
Skew / kurtosis-0.51 / -0.39
Normal?yes

Alignment

Alignment levelsample
Sample id availableyes
Samples80
Observations (total)240
Reps per samplemin 1 · mean 1 · max 1

Provenance & citation

ContributorEigenvector CORN
Origin · url [manual]https://www.eigenvector.com/data/
Origin · url [manual]https://www.eigenvector.com/data/Corn/
Origin · url [manual]https://eigenvector.com/wp-content/uploads/2019/06/corn.mat_.zip
Origin · url [manual]https://www.eigenvector.com/data/Corn/corn.mat
Origin · script [manual]source_to_standard.py — standardization script (maintainer-only)

Governance & integrity

Tierprivate
LicenseLicenseRef-not-cleared
Permitted useResearch and benchmarking; private use only.
Access policyManual download / private-use-only per source.
RedistributionEigenvector page says the data are available for download but no explicit redistribution license was detected.
Content version1.0.0
Schema / protocol2.0
Content hash8be9b3e72f065a78…
Processing hashb3772222aa96d782…
Metadata hashb9291350222a0b5c…

Load this dataset

# pip install nirs4all-datasets
from nirs4all_datasets import get

# private dataset — export requires a Dataverse token
ds = get("corn_eigenvector_nir", token="…")
X, y = ds.x(), ds.y()
print(X.shape, y.shape)

Metadata downloads are available for public datasets only. The dataset bytes are never served here — fetch them from the origin / DOI above.