← Back to the catalog
Private

Nirpy Research milk NIR classification

nirpy · NIR

Nirpy Research milk NIR classification. v2.0 standardized NIRS package: 1 spectral source(s), 1 declared target(s). Auto-generated from dataset_card.json (verify before publication).

nirv2nirpy
🔒
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.
450
samples
601
wavelengths
1
sources
1
targets
13
metadata
NIR
family

Dataset property explorer

Mean profile risk0.46
Highest axisArtefacts locaux · 1.00
Diagnostics8
Sources profiled1
Nirpy Research milk NIR classification property profile0.250.50.751integritynoiseartefactsbaselinePCA outliersreferencerepeatabilitystructureNirpy Research milk NIR classification profileintegrity: 0.00noise: 0.00artefacts: 1.00baseline: 1.00PCA outliers: 0.88reference: 0.05repeatability: 0.00structure: 0.73Nirpy Research …0 center · 1 outer ring · outward = stronger anomaly / heterogeneity signal

Profile axes

Intégrité0.00
Artefacts locaux1.00
Bruit0.00
Outliers PCA0.88
Distance à la référence0.05
Répétabilité0.00
Baseline / forme1.00
Structure multi-régimes0.73
Diagnostic hypotheses00.250.50.751hypothesis scoreErreur interpolation / réécha…Erreur interpolation / rééchantillonnage: 0.700.70Splice / raccord détecteursSplice / raccord détecteurs: 0.690.69Erreur calibration / référenc…Erreur calibration / référence blanche: 0.660.66Fond différentFond différent: 0.610.61Signature VERA25-likeSignature VERA25-like: 0.610.61Différence de sonde / géométr…Différence de sonde / géométrie: 0.530.53Spectre saturé / clippingSpectre saturé / clipping: 0.470.47Spectre hors domaine valideSpectre hors domaine valide: 0.460.46
DiagnosticScoreForceSignauxInterprétation probable
Erreur interpolation / rééchantillonnageX0.70moyenneSpike rate 1.00, Jump rate 1.00, SNR normal/élevé 1.00Artefacts numériques ou traitement spectral incorrect.
Splice / raccord détecteursX0.69moyenneSpike rate 1.00, Jump rate 1.00, SNR non dégradé 1.00Rupture aux jonctions de détecteurs, calibration locale ou sonde différente.
Erreur calibration / référence blancheX0.66moyenneBaseline/mean/area 1.00, artefacts locaux 1.00, Mahalanobis / T2 0.88Décalage systématique entre campagnes, instruments ou référence blanche.
Fond différentX0.61moyenneBaseline/mean/area 1.00, Mahalanobis / T2 0.88, PCA Q 0.72Effet systématique du support, blanc/noir, transflectance ou environnement de mesure.
Signature VERA25-likeX0.61moyenneSpike rate 1.00, Jump rate 1.00, Mahalanobis / T2 0.88Combinaison possible changement de sonde + splice, amplifiée par géométrie, fond ou calibration.
Différence de sonde / géométrieX0.53moyenneBaseline/mean/area 1.00, Mahalanobis / T2 0.88, PCA Q 0.72Modification de l'illumination, collecte, angle ou distance sonde-échantillon.
Spectre saturé / clippingX0.47moyenneBaseline/mean/area 1.00, Jump rate 1.00, PCA Q 0.72Détecteur saturé ou dynamique insuffisante.
Spectre hors domaine valideX0.46moyenneMahalanobis / T2 0.88, Structure PCA 0.73Variété, espèce, lot ou condition différente mais physiquement plausible.

Spectral sources

Nirpy Research milk NIR classification

X · NIR
Nirpy Research milk NIR classification spectra01230200400600800q05-q95 envelopeq25-q75 envelopemedian spectrummedianq25–q75q05–q95wavelength / index2index — median 2.316 (q25–q75 2.296–2.333)6index — median 2.29 (q25–q75 2.27–2.307)11index — median 2.231 (q25–q75 2.211–2.247)15index — median 2.155 (q25–q75 2.135–2.17)19index — median 2.015 (q25–q75 1.996–2.028)24index — median 1.807 (q25–q75 1.789–1.817)28index — median 1.679 (q25–q75 1.661–1.688)32index — median 1.602 (q25–q75 1.585–1.611)37index — median 1.557 (q25–q75 1.54–1.566)41index — median 1.54 (q25–q75 1.523–1.549)45index — median 1.525 (q25–q75 1.509–1.534)49index — median 1.514 (q25–q75 1.499–1.524)54index — median 1.507 (q25–q75 1.492–1.517)58index — median 1.505 (q25–q75 1.49–1.516)62index — median 1.507 (q25–q75 1.492–1.52)67index — median 1.513 (q25–q75 1.499–1.528)71index — median 1.52 (q25–q75 1.506–1.536)75index — median 1.528 (q25–q75 1.514–1.545)80index — median 1.532 (q25–q75 1.518–1.55)84index — median 1.524 (q25–q75 1.509–1.541)88index — median 1.51 (q25–q75 1.496–1.529)93index — median 1.49 (q25–q75 1.475–1.509)97index — median 1.459 (q25–q75 1.443–1.478)101index — median 1.41 (q25–q75 1.395–1.429)106index — median 1.332 (q25–q75 1.317–1.351)110index — median 1.256 (q25–q75 1.242–1.274)114index — median 1.171 (q25–q75 1.157–1.187)119index — median 1.057 (q25–q75 1.044–1.071)123index — median 0.9708 (q25–q75 0.9581–0.9844)127index — median 0.9001 (q25–q75 0.8875–0.9124)131index — median 0.8355 (q25–q75 0.8239–0.8468)136index — median 0.7472 (q25–q75 0.7364–0.757)140index — median 0.653 (q25–q75 0.6439–0.662)144index — median 0.5318 (q25–q75 0.5252–0.5393)149index — median 0.3777 (q25–q75 0.3734–0.383)153index — median 0.2857 (q25–q75 0.2824–0.2897)157index — median 0.2304 (q25–q75 0.2278–0.2341)162index — median 0.1959 (q25–q75 0.1937–0.199)166index — median 0.1806 (q25–q75 0.1786–0.1832)170index — median 0.1708 (q25–q75 0.169–0.1732)175index — median 0.1636 (q25–q75 0.162–0.1658)179index — median 0.1609 (q25–q75 0.1594–0.163)183index — median 0.1611 (q25–q75 0.1599–0.1633)188index — median 0.1668 (q25–q75 0.1656–0.169)192index — median 0.1754 (q25–q75 0.1742–0.1777)196index — median 0.1866 (q25–q75 0.1854–0.1888)201index — median 0.203 (q25–q75 0.2014–0.2049)205index — median 0.2176 (q25–q75 0.2157–0.2191)209index — median 0.2325 (q25–q75 0.2305–0.2342)214index — median 0.2519 (q25–q75 0.2497–0.2536)218index — median 0.2674 (q25–q75 0.265–0.2695)222index — median 0.2823 (q25–q75 0.2798–0.2847)226index — median 0.2988 (q25–q75 0.296–0.3019)231index — median 0.3205 (q25–q75 0.3173–0.3245)235index — median 0.3369 (q25–q75 0.3335–0.3413)239index — median 0.3527 (q25–q75 0.3488–0.3573)244index — median 0.3703 (q25–q75 0.3661–0.3758)248index — median 0.3827 (q25–q75 0.3782–0.3887)252index — median 0.3939 (q25–q75 0.3894–0.4002)257index — median 0.4068 (q25–q75 0.4019–0.4136)261index — median 0.4163 (q25–q75 0.4109–0.4233)265index — median 0.425 (q25–q75 0.4194–0.4322)270index — median 0.434 (q25–q75 0.4284–0.4417)274index — median 0.4399 (q25–q75 0.4339–0.4479)278index — median 0.4435 (q25–q75 0.4374–0.452)283index — median 0.4447 (q25–q75 0.4384–0.4531)287index — median 0.4437 (q25–q75 0.4374–0.4521)291index — median 0.4421 (q25–q75 0.4354–0.4504)296index — median 0.4377 (q25–q75 0.4311–0.4456)300index — median 0.4315 (q25–q75 0.425–0.4393)304index — median 0.4231 (q25–q75 0.4164–0.4308)308index — median 0.4127 (q25–q75 0.406–0.4202)313index — median 0.3973 (q25–q75 0.3907–0.4043)317index — median 0.3852 (q25–q75 0.3785–0.3921)321index — median 0.3749 (q25–q75 0.3683–0.3816)326index — median 0.3595 (q25–q75 0.353–0.3657)330index — median 0.3448 (q25–q75 0.3384–0.3506)334index — median 0.3313 (q25–q75 0.3248–0.3367)339index — median 0.3186 (q25–q75 0.3124–0.3239)343index — median 0.3114 (q25–q75 0.3051–0.3166)347index — median 0.3057 (q25–q75 0.2994–0.3108)352index — median 0.3005 (q25–q75 0.294–0.3055)356index — median 0.2982 (q25–q75 0.2918–0.3032)360index — median 0.2962 (q25–q75 0.2895–0.301)365index — median 0.2912 (q25–q75 0.2845–0.2959)369index — median 0.2832 (q25–q75 0.2767–0.2876)373index — median 0.2704 (q25–q75 0.2639–0.2745)378index — median 0.2443 (q25–q75 0.238–0.2477)382index — median 0.2128 (q25–q75 0.2071–0.2156)386index — median 0.1731 (q25–q75 0.1688–0.1752)390index — median 0.1328 (q25–q75 0.1297–0.1344)395index — median 0.09227 (q25–q75 0.09062–0.09325)399index — median 0.07129 (q25–q75 0.07019–0.07207)403index — median 0.05971 (q25–q75 0.05897–0.06045)408index — median 0.05318 (q25–q75 0.05256–0.05386)412index — median 0.05083 (q25–q75 0.05024–0.05152)416index — median 0.04963 (q25–q75 0.0491–0.05042)421index — median 0.04959 (q25–q75 0.04899–0.05038)425index — median 0.05123 (q25–q75 0.05048–0.05194)429index — median 0.05322 (q25–q75 0.0525–0.05407)434index — median 0.05616 (q25–q75 0.05542–0.05703)438index — median 0.05913 (q25–q75 0.05827–0.05992)442index — median 0.06208 (q25–q75 0.0612–0.06289)447index — median 0.06586 (q25–q75 0.06493–0.06681)451index — median 0.06846 (q25–q75 0.06742–0.06952)455index — median 0.07038 (q25–q75 0.0692–0.07135)460index — median 0.07172 (q25–q75 0.07028–0.07277)464index — median 0.07261 (q25–q75 0.07104–0.07375)468index — median 0.07391 (q25–q75 0.07222–0.07506)473index — median 0.0764 (q25–q75 0.07445–0.0777)477index — median 0.07798 (q25–q75 0.07587–0.07931)481index — median 0.07854 (q25–q75 0.07635–0.08007)485index — median 0.0793 (q25–q75 0.07703–0.08108)490index — median 0.08314 (q25–q75 0.08077–0.08518)494index — median 0.08784 (q25–q75 0.08518–0.08994)498index — median 0.09236 (q25–q75 0.08922–0.09436)503index — median 0.09695 (q25–q75 0.09357–0.09899)507index — median 0.09998 (q25–q75 0.09651–0.1023)511index — median 0.1025 (q25–q75 0.09889–0.105)516index — median 0.1056 (q25–q75 0.1019–0.1083)520index — median 0.1084 (q25–q75 0.1044–0.1113)524index — median 0.1114 (q25–q75 0.1073–0.1141)529index — median 0.1145 (q25–q75 0.1106–0.1176)533index — median 0.1167 (q25–q75 0.1127–0.1197)537index — median 0.1175 (q25–q75 0.1134–0.1203)542index — median 0.1162 (q25–q75 0.1117–0.1191)546index — median 0.1135 (q25–q75 0.1089–0.1167)550index — median 0.1111 (q25–q75 0.1062–0.1142)555index — median 0.1097 (q25–q75 0.1051–0.113)559index — median 0.11 (q25–q75 0.1056–0.1134)563index — median 0.1108 (q25–q75 0.1064–0.114)567index — median 0.1114 (q25–q75 0.1071–0.1146)572index — median 0.1123 (q25–q75 0.108–0.1154)576index — median 0.1123 (q25–q75 0.1084–0.1156)580index — median 0.1121 (q25–q75 0.1085–0.1154)585index — median 0.1111 (q25–q75 0.1073–0.1141)589index — median 0.1103 (q25–q75 0.1067–0.1133)593index — median 0.1128 (q25–q75 0.1092–0.1155)598index — median 0.125 (q25–q75 0.122–0.128)602index — median 0.145 (q25–q75 0.1416–0.1478)

Sampling

Wavelengths601
Axis range2–602 index
Mean spacing1 index
Griduniform
Observations450

Signal & quality

Value range0.0474 – 2.43
Mean range0.0499 – 2.32
Mean level0.5091
Area304.7
PTP2.271
Noise RMS7.9432e-05
SNR6.4e+03
SNR dB8e+01 dB
Dynamic range2.27
Smoothness0.0007659
Saturated0.0%
X-outliers118

Integrity & artefacts

NaN ratio0.00%
Inf count0
Zero ratio0.00%
Spike count17,195
Spike rate6.38%
Jump count12,825
Jump rate4.75%
Clip fraction0.00%

Shape & reference

Baseline slope-1.6097
Curvature RMS0.00076631
D1 RMS0.0094273
RMS to mean0.0096895
RMS p950.025556
SAM to mean0.0052404
SAM p950.012601
Affine offset p950.010393
Affine gain p95 Δ0.037651
Affine residual p950.0076112
Xcorr lag p950

Outliers & repeatability

PCA Q p95/median5.8
Hotelling T2 p95/median7.1
Mahalanobis H p95/median2.7
Repeat groups0

Dimensionality (PCA)

Effective rank1.7
PCs → 95% var2
PCs → 99% var3
Top-10 cum. var99.9%
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_reflectance0.509071.00fortValeur 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_curve304.721.00fortValeur 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_peak2.27120.00faibleVariabilité forteSaturationmax(mean_spectrum) - min(mean_spectrum)alert increases when dynamic range is abnormally flat
Amplitude globaleVarianceamplitude.variance0.340860.00faibleNormal ou hétérogèneMauvais contactvar(X finite)alert increases when variance/dynamic range is abnormally flat
BruitNoise RMSnoise.noise_rms7.9432e-050.00faibleStableLampe, détecteurmedian MAD(second derivative) * 1.4826 / sqrt(6)alert = noise_rms / signal_scale, saturated at 5%
BruitSNRnoise.snr6408.80.00faibleBon signalAcquisitionmean(abs(X)) / noise_rmsalert decreases with SNR dB; >=40 dB is treated as low alert
BruitBandwise SNRnoise.bandwise_snr_min1721.70.00faibleZone 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_count17,1951.00fortArtefactsCosmic rays, splicecount robust outliers in second derivativealert follows spike_rate, saturated at 1%
Artefacts locauxSpike rateartefacts.spike_rate6.38%1.00fortSpectre suspectInterpolationspike_count / (n_samples * (n_features - 2))alert = min(1, spike_rate / 0.01)
Artefacts locauxJump countartefacts.jump_count12,8251.00fortRaccord détecteurSplicecount robust outliers in first derivativealert follows jump_rate, saturated at 1%
Artefacts locauxJump rateartefacts.jump_rate4.75%1.00fortProblème spectralCalibrationjump_count / (n_samples * (n_features - 1))alert = min(1, jump_rate / 0.01)
Artefacts locauxClip fractionartefacts.clip_fraction0.00074%0.00faibleNormalDétecteur saturéfraction of finite cells equal to repeated min/max extremaalert = min(1, clip_fraction / 0.01)
Forme spectraleBaseline slopeshape.baseline_slope-1.60971.00fortDériveÉclairementlinear slope of mean_spectrum over normalized axisalert = abs(slope / signal_scale), saturated at 0.5
Forme spectraleCurvature RMSshape.curvature_rms0.000766310.03faibleLisseFond, splicemedian RMS(second derivative per spectrum)alert = curvature_rms / signal_scale, saturated at 1%
Forme spectraleD1 RMSshape.d1_rms0.00942730.08faiblePlatBiologie ou artefactmedian RMS(first derivative per spectrum)alert = d1_rms / signal_scale, saturated at 5%
Outliers multivariésPCA Q (SPE)outliers.pca_q_ratio5.79820.72fortSpectre atypiqueArtefact, mélangep95(Q/SPE residual) / median(Q/SPE residual)alert = min(1, pca_q_ratio / 8)
Outliers multivariésHotelling T²outliers.hotelling_t2_ratio7.07590.88fortExtrême mais cohérentVariabilité naturellep95(Hotelling T2) / median(Hotelling T2)alert = min(1, hotelling_t2_ratio / 8)
Outliers multivariésMahalanobis Houtliers.mahalanobis_h_ratio2.66010.67moyenOutlier globalDomaine différentp95(sqrt(T2)) / median(sqrt(T2))alert = min(1, mahalanobis_h_ratio / 4)
Comparaison à référenceRMS to mean spectrumreference.rms_to_mean_spectrum_p950.0255560.05faibleTypiqueDomain 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_p950.0126010.04faibleSimilaireFond, 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_density53.7320.73fortSous-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_p952.1870.59moyenSpectre 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_p950.591440.73fortSpectre atypiqueDiverses causesp95 IsolationForest anomaly score on PCA scoresalert follows structure complexity; raw score is implementation-dependent
X PCA score plot-0.50.00.51.0-0.4-0.20.00.20.4PC1 0.4947 · PC2 0.2277PC1 0.4908 · PC2 0.2343PC1 0.4935 · PC2 0.2282PC1 0.5606 · PC2 0.2236PC1 0.5918 · PC2 0.2279PC1 0.6076 · PC2 0.2182PC1 0.6049 · PC2 0.2223PC1 0.6922 · PC2 0.2394PC1 0.6774 · PC2 0.2357PC1 0.6613 · PC2 0.2293PC1 0.6657 · PC2 0.2275PC1 0.6205 · PC2 0.2337PC1 0.5983 · PC2 0.23PC1 0.5764 · PC2 0.2405PC1 0.5454 · PC2 0.2315PC1 0.4937 · PC2 0.2307PC1 0.4998 · PC2 0.2268PC1 0.4825 · PC2 0.2088PC1 0.4495 · PC2 0.2275PC1 0.4655 · PC2 0.2322PC1 0.4834 · PC2 0.2313PC1 0.4786 · PC2 0.2285PC1 0.5052 · PC2 0.2286PC1 0.548 · PC2 0.233PC1 0.5987 · PC2 0.2387PC1 0.6141 · PC2 0.2275PC1 0.6325 · PC2 0.2347PC1 0.6071 · PC2 0.2404PC1 0.579 · PC2 0.2269PC1 0.5728 · PC2 0.2218PC1 0.5552 · PC2 0.2164PC1 0.5256 · PC2 0.2161PC1 0.4652 · PC2 0.223PC1 0.4706 · PC2 0.2279PC1 0.4458 · PC2 0.2235PC1 0.4655 · PC2 0.2298PC1 0.4977 · PC2 0.2285PC1 0.5381 · PC2 0.2258PC1 0.5655 · PC2 0.2369PC1 0.5771 · PC2 0.2291PC1 0.604 · PC2 0.2388PC1 0.6345 · PC2 0.2339PC1 0.6324 · PC2 0.2336PC1 0.6028 · PC2 0.2425PC1 0.6055 · PC2 0.2342PC1 0.5844 · PC2 0.2389PC1 0.5986 · PC2 0.2392PC1 0.614 · PC2 0.2272PC1 0.6142 · PC2 0.2349PC1 0.5796 · PC2 0.2409PC1 -0.4042 · PC2 0.1252PC1 -0.4114 · PC2 0.1309PC1 -0.4066 · PC2 0.1353PC1 -0.381 · PC2 0.1307PC1 -0.2808 · PC2 0.1215PC1 -0.1924 · PC2 0.1008PC1 -0.252 · PC2 0.08189PC1 -0.2973 · PC2 0.09892PC1 -0.3262 · PC2 0.07407PC1 -0.3939 · PC2 0.1222PC1 -0.4332 · PC2 0.1234PC1 -0.4351 · PC2 0.1204PC1 -0.4395 · PC2 0.111PC1 -0.4203 · PC2 0.1009PC1 -0.4235 · PC2 0.09667PC1 -0.4206 · PC2 0.1134PC1 -0.3986 · PC2 0.1228PC1 -0.327 · PC2 0.1143PC1 -0.3064 · PC2 0.1308PC1 -0.2962 · PC2 0.1345PC1 -0.2838 · PC2 0.1298PC1 -0.3087 · PC2 0.0953PC1 -0.305 · PC2 0.04501PC1 -0.4024 · PC2 0.09914PC1 -0.4728 · PC2 0.112PC1 -0.4897 · PC2 0.1052PC1 -0.442 · PC2 0.1023PC1 -0.4407 · PC2 0.1293PC1 -0.3853 · PC2 0.1318PC1 -0.3816 · PC2 0.1267PC1 -0.4433 · PC2 0.1256PC1 -0.4569 · PC2 0.1333PC1 -0.4438 · PC2 0.1348PC1 -0.4225 · PC2 0.122PC1 -0.4487 · PC2 0.1245PC1 -0.4347 · PC2 0.102PC1 -0.4561 · PC2 0.1122PC1 -0.469 · PC2 0.11PC1 -0.4596 · PC2 0.1073PC1 -0.4034 · PC2 0.1089PC1 -0.388 · PC2 0.1198PC1 -0.3315 · PC2 0.1153PC1 -0.4084 · PC2 0.04769PC1 -0.4687 · PC2 0.1024PC1 -0.4478 · PC2 0.1259PC1 -0.4582 · PC2 0.1268PC1 -0.4702 · PC2 0.1321PC1 -0.4807 · PC2 0.1335PC1 -0.4777 · PC2 0.1357PC1 -0.4773 · PC2 0.1361PC1 -0.05637 · PC2 0.08157PC1 -0.06785 · PC2 0.09414PC1 -0.04943 · PC2 0.09287PC1 0.02368 · PC2 0.05895PC1 0.04253 · PC2 0.07088PC1 0.05078 · PC2 0.09142PC1 0.04835 · PC2 0.09075PC1 0.02876 · PC2 0.0969PC1 -0.05743 · PC2 0.09917PC1 -0.07653 · PC2 0.09375PC1 -0.1648 · PC2 0.1028PC1 -0.1519 · PC2 0.09646PC1 -0.1107 · PC2 0.07445PC1 -0.09098 · PC2 0.08239PC1 -0.05545 · PC2 0.07968PC1 -0.04496 · PC2 0.1009PC1 0.02567 · PC2 0.1009PC1 0.0506 · PC2 0.08622PC1 0.02785 · PC2 0.1161PC1 -0.02696 · PC2 0.1056PC1 -0.0583 · PC2 0.07776PC1 -0.1185 · PC2 0.09627PC1 -0.184 · PC2 0.09341PC1 -0.1459 · PC2 0.08807PC1 -0.0543 · PC2 0.07862PC1 -0.01758 · PC2 0.08651PC1 -0.02552 · PC2 0.09897PC1 0.004492 · PC2 0.09326PC1 -0.01954 · PC2 0.1086PC1 -0.03874 · PC2 0.08422PC1 -0.06001 · PC2 0.06761PC1 -0.1186 · PC2 0.08217PC1 -0.1212 · PC2 0.09152PC1 -0.1352 · PC2 0.09426PC1 -0.1958 · PC2 0.09376PC1 -0.1874 · PC2 0.0956PC1 -0.1798 · PC2 0.09238PC1 -0.1529 · PC2 0.09441PC1 -0.08789 · PC2 0.08288PC1 -0.03795 · PC2 0.08499PC1 -0.009116 · PC2 0.1078PC1 0.03648 · PC2 0.08416PC1 -0.005737 · PC2 0.0984PC1 -0.07208 · PC2 0.0601PC1 -0.1087 · PC2 0.05724PC1 -0.03571 · PC2 0.06643PC1 -0.05629 · PC2 0.08717PC1 -0.04895 · PC2 0.08633PC1 -0.06849 · PC2 0.1007PC1 -0.0793 · PC2 0.1021PC1 -0.356 · PC2 0.06564PC1 -0.3636 · PC2 0.07338PC1 -0.3482 · PC2 0.07059PC1 -0.3139 · PC2 0.06672PC1 -0.2618 · PC2 0.06176PC1 -0.23 · PC2 0.05908PC1 -0.2476 · PC2 0.06072PC1 -0.2809 · PC2 0.06794PC1 -0.2769 · PC2 0.0159PC1 -0.3853 · PC2 0.04191PC1 -0.426 · PC2 0.06411PC1 -0.4327 · PC2 0.04426PC1 -0.3821 · PC2 0.03456PC1 -0.35 · PC2 0.0258PC1 -0.3416 · PC2 0.04237PC1 -0.2852 · PC2 0.03965PC1 -0.2597 · PC2 0.06151PC1 -0.2174 · PC2 0.069PC1 -0.1734 · PC2 0.06394PC1 -0.1134 · PC2 0.04964PC1 -0.1154 · PC2 0.05034PC1 -0.1251 · PC2 0.05775PC1 -0.18 · PC2 0.07947PC1 -0.2255 · PC2 0.07525PC1 -0.2987 · PC2 0.07184PC1 -0.3343 · PC2 0.06085PC1 -0.3447 · PC2 0.0535PC1 -0.2876 · PC2 0.06195PC1 -0.2044 · PC2 0.07028PC1 -0.2052 · PC2 0.0736PC1 -0.2582 · PC2 0.05547PC1 -0.341 · PC2 0.06702PC1 -0.3776 · PC2 0.05433PC1 -0.3828 · PC2 0.06086PC1 -0.3451 · PC2 0.04615PC1 -0.2676 · PC2 0.05847PC1 -0.2338 · PC2 0.06836PC1 -0.1963 · PC2 0.07208PC1 -0.1602 · PC2 0.04714PC1 -0.2538 · PC2 0.05526PC1 -0.3202 · PC2 0.05005PC1 -0.3115 · PC2 0.05962PC1 -0.2559 · PC2 0.06527PC1 -0.1544 · PC2 0.04759PC1 -0.1569 · PC2 0.03755PC1 -0.2362 · PC2 0.04603PC1 -0.3317 · PC2 0.04953PC1 -0.3591 · PC2 0.04296PC1 -0.3634 · PC2 0.05035PC1 -0.3642 · PC2 0.05292PC1 -0.04419 · PC2 -0.001568PC1 -0.03793 · PC2 -0.0003776PC1 -0.02499 · PC2 0.0005221PC1 -0.02632 · PC2 -0.004104PC1 -0.08483 · PC2 -0.00607PC1 -0.1285 · PC2 0.009043PC1 -0.02296 · PC2 0.02313PC1 0.1464 · PC2 0.03906PC1 0.1799 · PC2 0.02533PC1 0.2284 · PC2 -0.002311PC1 0.1832 · PC2 -0.02654PC1 0.1637 · PC2 -0.02705PC1 0.2406 · PC2 -0.003799PC1 0.3707 · PC2 0.0009947PC1 0.3128 · PC2 0.01704PC1 0.09751 · PC2 0.005375PC1 -0.05492 · PC2 0.005412PC1 -0.1602 · PC2 0.024PC1 -0.1685 · PC2 0.03077PC1 -0.2001 · PC2 -0.001751PC1 -0.1454 · PC2 -0.001952PC1 0.01211 · PC2 -0.009533PC1 0.153 · PC2 0.005344PC1 0.1218 · PC2 -0.03982PC1 -0.03729 · PC2 -0.0144PC1 -0.05037 · PC2 -0.0006388PC1 0.07471 · PC2 -0.02774PC1 -0.01019 · PC2 -0.01718PC1 -0.0523 · PC2 0.000335PC1 -0.1571 · PC2 0.006525PC1 0.01293 · PC2 0.007741PC1 -0.04095 · PC2 -0.01118PC1 0.113 · PC2 0.003148PC1 0.05481 · PC2 -0.004478PC1 0.03598 · PC2 -0.0206PC1 0.002104 · PC2 0.003856PC1 0.0128 · PC2 0.01182PC1 0.04412 · PC2 0.008606PC1 0.116 · PC2 0.03814PC1 0.1676 · PC2 0.04217PC1 0.1725 · PC2 0.03059PC1 0.09492 · PC2 0.03727PC1 0.07373 · PC2 -0.004153PC1 0.1189 · PC2 -0.05922PC1 0.09604 · PC2 -0.02242PC1 0.04872 · PC2 0.003443PC1 0.0229 · PC2 0.01665PC1 0.003441 · PC2 0.0272PC1 0.00842 · PC2 0.02967PC1 -0.004475 · PC2 0.02815PC1 -0.2307 · PC2 -0.07625PC1 -0.2454 · PC2 -0.05854PC1 -0.2544 · PC2 -0.05468PC1 -0.2339 · PC2 -0.04608PC1 -0.2243 · PC2 -0.04038PC1 -0.2224 · PC2 -0.04805PC1 -0.2677 · PC2 -0.05501PC1 -0.2928 · PC2 -0.05474PC1 -0.2647 · PC2 -0.09857PC1 -0.1817 · PC2 -0.1138PC1 -0.2335 · PC2 -0.08906PC1 -0.1932 · PC2 -0.06121PC1 -0.1636 · PC2 -0.047PC1 -0.1281 · PC2 -0.051PC1 -0.133 · PC2 -0.02984PC1 -0.1266 · PC2 -0.02457PC1 -0.1259 · PC2 -0.05548PC1 -0.1499 · PC2 -0.08654PC1 -0.2293 · PC2 -0.05566PC1 -0.2972 · PC2 -0.04112PC1 -0.3205 · PC2 -0.03877PC1 -0.2922 · PC2 -0.04365PC1 -0.2213 · PC2 -0.08804PC1 -0.1726 · PC2 -0.05132PC1 -0.09597 · PC2 -0.06528PC1 -0.06898 · PC2 -0.04962PC1 -0.02169 · PC2 -0.07817PC1 -0.0632 · PC2 -0.07853PC1 -0.1842 · PC2 -0.05217PC1 -0.2517 · PC2 -0.05155PC1 -0.2524 · PC2 -0.06078PC1 -0.3262 · PC2 -0.05582PC1 -0.3342 · PC2 -0.06223PC1 -0.3287 · PC2 -0.06585PC1 -0.3137 · PC2 -0.04777PC1 -0.2478 · PC2 -0.04854PC1 -0.1919 · PC2 -0.04894PC1 -0.1916 · PC2 -0.0409PC1 -0.2146 · PC2 -0.06154PC1 -0.2843 · PC2 -0.04254PC1 -0.3785 · PC2 -0.0511PC1 -0.2433 · PC2 -0.05025PC1 -0.2019 · PC2 -0.05597PC1 -0.1141 · PC2 -0.06746PC1 -0.1186 · PC2 -0.06409PC1 -0.1327 · PC2 -0.05689PC1 -0.1289 · PC2 -0.05655PC1 -0.137 · PC2 -0.05722PC1 -0.1483 · PC2 -0.05784PC1 -0.1537 · PC2 -0.05784PC1 -0.08462 · PC2 -0.1004PC1 -0.09261 · PC2 -0.07814PC1 -0.1002 · PC2 -0.06846PC1 -0.1049 · PC2 -0.06769PC1 -0.07218 · PC2 -0.07108PC1 -0.05381 · PC2 -0.06841PC1 0.05768 · PC2 -0.09416PC1 0.08789 · PC2 -0.06916PC1 -0.01313 · PC2 -0.104PC1 -0.1262 · PC2 -0.09101PC1 -0.1726 · PC2 -0.08581PC1 -0.2005 · PC2 -0.0753PC1 -0.1536 · PC2 -0.09199PC1 -0.09449 · PC2 -0.09346PC1 -0.04427 · PC2 -0.083PC1 0.01482 · PC2 -0.08554PC1 0.02461 · PC2 -0.07155PC1 0.042 · PC2 -0.07024PC1 0.01941 · PC2 -0.06842PC1 -0.00535 · PC2 -0.07897PC1 -0.07562 · PC2 -0.09068PC1 -0.1432 · PC2 -0.07322PC1 -0.1403 · PC2 -0.08245PC1 -0.06462 · PC2 -0.08663PC1 -0.03724 · PC2 -0.0759PC1 -0.01791 · PC2 -0.07001PC1 0.02627 · PC2 -0.06539PC1 0.0692 · PC2 -0.06516PC1 0.01861 · PC2 -0.0521PC1 -0.05611 · PC2 -0.05908PC1 -0.1261 · PC2 -0.06691PC1 -0.1327 · PC2 -0.06108PC1 -0.1758 · PC2 -0.08712PC1 -0.1728 · PC2 -0.0837PC1 -0.1671 · PC2 -0.1057PC1 -0.06504 · PC2 -0.1087PC1 -0.0005452 · PC2 -0.08125PC1 -0.01235 · PC2 -0.0761PC1 0.005799 · PC2 -0.06123PC1 0.04346 · PC2 -0.06356PC1 0.05229 · PC2 -0.0659PC1 0.03832 · PC2 -0.0743PC1 0.06283 · PC2 -0.0958PC1 -0.001402 · PC2 -0.08906PC1 -0.03718 · PC2 -0.08566PC1 -0.01845 · PC2 -0.0637PC1 -0.01565 · PC2 -0.06PC1 -0.01874 · PC2 -0.05949PC1 -0.01988 · PC2 -0.05316PC1 -0.02458 · PC2 -0.05279PC1 0.1384 · PC2 -0.133PC1 0.1331 · PC2 -0.1188PC1 0.1439 · PC2 -0.124PC1 0.1671 · PC2 -0.1163PC1 0.2168 · PC2 -0.1187PC1 0.2012 · PC2 -0.1114PC1 0.1875 · PC2 -0.09365PC1 0.1763 · PC2 -0.09622PC1 0.1606 · PC2 -0.11PC1 0.1476 · PC2 -0.1407PC1 0.05573 · PC2 -0.1094PC1 0.03252 · PC2 -0.1098PC1 0.02259 · PC2 -0.1058PC1 0.04003 · PC2 -0.1078PC1 0.1088 · PC2 -0.1148PC1 0.1478 · PC2 -0.09975PC1 0.1898 · PC2 -0.1144PC1 0.2499 · PC2 -0.1035PC1 0.2689 · PC2 -0.1136PC1 0.2024 · PC2 -0.1211PC1 0.1334 · PC2 -0.115PC1 0.08015 · PC2 -0.1223PC1 0.07927 · PC2 -0.1325PC1 0.1086 · PC2 -0.1278PC1 0.1917 · PC2 -0.1382PC1 0.2265 · PC2 -0.1236PC1 0.2431 · PC2 -0.1208PC1 0.154 · PC2 -0.1105PC1 0.1206 · PC2 -0.1168PC1 0.1161 · PC2 -0.1561PC1 0.09072 · PC2 -0.1245PC1 0.1286 · PC2 -0.1541PC1 0.2236 · PC2 -0.1618PC1 0.1017 · PC2 -0.1267PC1 0.05789 · PC2 -0.1117PC1 0.06181 · PC2 -0.1116PC1 0.1295 · PC2 -0.1144PC1 0.2182 · PC2 -0.1033PC1 0.2356 · PC2 -0.1173PC1 0.1307 · PC2 -0.1082PC1 0.09069 · PC2 -0.1119PC1 0.118 · PC2 -0.1246PC1 0.1722 · PC2 -0.1253PC1 0.1829 · PC2 -0.1111PC1 0.2064 · PC2 -0.134PC1 0.195 · PC2 -0.1529PC1 0.1873 · PC2 -0.1274PC1 0.1549 · PC2 -0.1238PC1 0.1508 · PC2 -0.1113PC1 0.1463 · PC2 -0.1127PC1 0.1823 · PC2 -0.2275PC1 0.1911 · PC2 -0.2206PC1 0.1952 · PC2 -0.2199PC1 0.1826 · PC2 -0.2214PC1 0.1825 · PC2 -0.2216PC1 0.1878 · PC2 -0.2205PC1 0.1937 · PC2 -0.2186PC1 0.1914 · PC2 -0.2162PC1 0.2107 · PC2 -0.2119PC1 0.2663 · PC2 -0.2281PC1 0.3414 · PC2 -0.2453PC1 0.3761 · PC2 -0.2481PC1 0.3895 · PC2 -0.2704PC1 0.3846 · PC2 -0.2624PC1 0.3623 · PC2 -0.2604PC1 0.3441 · PC2 -0.2733PC1 0.3384 · PC2 -0.2854PC1 0.3531 · PC2 -0.2933PC1 0.3335 · PC2 -0.3112PC1 0.2864 · PC2 -0.2945PC1 0.3022 · PC2 -0.2943PC1 0.3056 · PC2 -0.2963PC1 0.3007 · PC2 -0.2945PC1 0.2942 · PC2 -0.297PC1 0.2596 · PC2 -0.2829PC1 0.2218 · PC2 -0.2768PC1 0.2188 · PC2 -0.2713PC1 0.195 · PC2 -0.2592PC1 0.2064 · PC2 -0.2614PC1 0.1999 · PC2 -0.2269PC1 0.1731 · PC2 -0.2241PC1 0.1753 · PC2 -0.2114PC1 0.1655 · PC2 -0.209PC1 0.2148 · PC2 -0.2283PC1 0.2763 · PC2 -0.2335PC1 0.3074 · PC2 -0.2437PC1 0.2846 · PC2 -0.2266PC1 0.2036 · PC2 -0.2055PC1 0.179 · PC2 -0.2063PC1 0.1816 · PC2 -0.2139PC1 0.1732 · PC2 -0.2192PC1 0.1827 · PC2 -0.2083PC1 0.1946 · PC2 -0.2108PC1 0.2163 · PC2 -0.2124PC1 0.2644 · PC2 -0.2165PC1 0.2525 · PC2 -0.2126PC1 0.1711 · PC2 -0.1662PC1 0.1592 · PC2 -0.1853PC1 0.1817 · PC2 -0.204PC1 0.2093 · PC2 -0.205PC1 (81.0%)PC2 (17.6%)450 scores
PCA explained variance0%25%50%75%100%PC1: 81.0% (cumulative 81.0%)1PC2: 17.6% (cumulative 98.6%)2PC3: 0.9% (cumulative 99.5%)3PC4: 0.2% (cumulative 99.7%)4PC5: 0.1% (cumulative 99.8%)5PC6: 0.0% (cumulative 99.8%)6PC7: 0.0% (cumulative 99.8%)7PC8: 0.0% (cumulative 99.8%)8PC9: 0.0% (cumulative 99.8%)9PC10: 0.0% (cumulative 99.9%)10cumulative explained variancePC variancecumulativeprincipal component · cumulative (dashed)

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 1

labels

target · categorical
labels classes11: 505022: 505033: 505044: 505055: 505066: 505077: 505088: 505099: 5050
n / missing450 / 0
Classes9
Balance (entropy)1
Imbalance ratio1
Top class1 (50)

Metadata 1

original_sample_id

metadata · categorical
original_sample_id classes1/02/20181/02/2018: 991/02/2018.11/02/2018.1: 991/02/2018.21/02/2018.2: 991/02/2018.31/02/2018.3: 991/02/2018.41/02/2018.4: 991/02/2018.51/02/2018.5: 991/02/2018.61/02/2018.6: 991/02/2018.71/02/2018.7: 991/02/2018.81/02/2018.8: 991/02/2018.91/02/2018.9: 99+10 more+10 more: 9090
n / missing450 / 0
Classes50
Balance (entropy)1
Imbalance ratio1
Top class1/02/2018 (9)
Constant metadata 12
  • repository_urlhttps://github.com/nevernervous78/nirpyresearch
  • licenseMIT License
  • license_urlhttps://raw.githubusercontent.com/nevernervous78/nirpyresearch/master/LICENSE
  • rights_statusexplicit_open
  • usage_scopepublic_reuse_possible
  • public_release_allowedTrue
  • spectroscopy_typeNIR
  • axis_unitindex
  • axis_min2
  • axis_max602
  • n_points_original601
  • notesConverted from Nirpy Research GitHub raw data without interpolation or resampling.

Alignment

Alignment levelobservation
Sample id availableno
Samples450
Observations (total)450
Reps per samplemin 1 · mean 1 · max 1

Provenance & citation

ContributorNirpy Research GitHub data
Origin · url [open]https://github.com/nevernervous78/nirpyresearch/blob/master/data/milk.csv — GitHub
Origin · url [open]https://raw.githubusercontent.com/nevernervous78/nirpyresearch/master/LICENSE
Origin · script [manual]source_to_standard.py — standardization script (maintainer-only)

Governance & integrity

Tierprivate
LicenseLicenseRef-not-cleared
Permitted useResearch and benchmarking.
Access policyManual download / private-use-only per source.
RedistributionMIT license permits use, copy, modification, publication, distribution, sublicensing and sale with copyright/license notice.
Content version1.0.0
Schema / protocol2.0
Content hashab0456a6eccef9ca…
Processing hashd51b01bd92958d95…
Metadata hashfab747e17e9c0fcb…

Load this dataset

# pip install nirs4all-datasets
from nirs4all_datasets import get

# private dataset — export requires a Dataverse token
ds = get("nirpy_milk_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.