← Back to the catalog
Publicraw reproducible3 sources

Zenodo tomato leaf stress Vis-NIR matio

zenodo · NIR

Zenodo tomato leaf stress Vis-NIR matio. v2.0 standardized NIRS package: 3 spectral source(s), 3 declared target(s). Auto-generated from dataset_card.json (verify before publication).

nirv2zenodo
216
samples
2,496
wavelengths
3
sources
3
targets
16
metadata
NIR
family

Dataset property explorer

Mean profile risk0.53
Highest axisArtefacts locaux · 1.00
Diagnostics8
Sources profiled3
Zenodo tomato leaf stress Vis-NIR matio property profile0.250.50.751integritynoiseartefactsbaselinePCA outliersreferencerepeatabilitystructureZenodo tomato leaf stress Vis-NIR matio profileintegrity: 0.00noise: 0.01artefacts: 1.00baseline: 1.00PCA outliers: 0.51reference: 0.71repeatability: 0.00structure: 1.00Zenodo tomato l…0 center · 1 outer ring · outward = stronger anomaly / heterogeneity signal

Profile axes

Intégrité0.00
Artefacts locaux1.00
Bruit0.01
Outliers PCA0.51
Distance à la référence0.71
Répétabilité0.00
Baseline / forme1.00
Structure multi-régimes1.00
Diagnostic hypotheses00.250.50.751hypothesis scoreSplice / raccord détecteursSplice / raccord détecteurs: 0.760.76Erreur calibration / référenc…Erreur calibration / référence blanche: 0.670.67Erreur interpolation / réécha…Erreur interpolation / rééchantillonnage: 0.610.61Fond différentFond différent: 0.600.60Signature VERA25-likeSignature VERA25-like: 0.560.56Différence de sonde / géométr…Différence de sonde / géométrie: 0.520.52Spectre hors domaine valideSpectre hors domaine valide: 0.500.50Dataset multi-régimesDataset multi-régimes: 0.490.49
DiagnosticScoreForceSignauxInterprétation probable
Splice / raccord détecteursX_reflectance0.76forteSpike rate 1.00, Jump rate 1.00, SNR non dégradé 1.00Rupture aux jonctions de détecteurs, calibration locale ou sonde différente.
Erreur calibration / référence blancheX_transmittance0.67moyenneBaseline/mean/area 1.00, artefacts locaux 1.00, RMS/SAM référence 0.74Décalage systématique entre campagnes, instruments ou référence blanche.
Erreur interpolation / rééchantillonnageX_reflectance0.61moyenneSpike rate 1.00, Jump rate 1.00, SNR normal/élevé 1.00Artefacts numériques ou traitement spectral incorrect.
Fond différentX_transmittance0.60moyenneBaseline/mean/area 1.00, RMS/SAM référence 0.74, Mahalanobis / T2 0.53Effet systématique du support, blanc/noir, transflectance ou environnement de mesure.
Signature VERA25-likeX_reflectance0.56moyenneSpike rate 1.00, Jump rate 1.00, RMS/SAM référence 0.70Combinaison possible changement de sonde + splice, amplifiée par géométrie, fond ou calibration.
Différence de sonde / géométrieX_transmittance0.52moyenneBaseline/mean/area 1.00, RMS/SAM référence 0.74, Mahalanobis / T2 0.53Modification de l'illumination, collecte, angle ou distance sonde-échantillon.
Spectre hors domaine valideX_transmittance0.50moyenneStructure PCA 1.00, RMS/SAM référence 0.74, Mahalanobis / T2 0.53Variété, espèce, lot ou condition différente mais physiquement plausible.
Dataset multi-régimesX_transmittance0.49moyenneStructure PCA 1.00, RMS/SAM référence 0.74, Mahalanobis / T2 0.53Mélange de campagnes, opérateurs, lots, setups ou sous-populations spectrales.

Spectral sources

absorbance

X_absorbance · NIR
absorbance spectra-0.50.00.51.02004006008001,0001,200q05-q95 envelopeq25-q75 envelopemedian spectrummedianq25–q75q05–q95wavelength / nm394.92nm — median 0.9786 (q25–q75 0.9725–0.9849)399.62nm — median 0.9247 (q25–q75 0.915–0.9342)404.31nm — median 0.9288 (q25–q75 0.9187–0.9376)409.01nm — median 0.9331 (q25–q75 0.9239–0.9413)413.7nm — median 0.9366 (q25–q75 0.927–0.9441)418.39nm — median 0.9396 (q25–q75 0.9302–0.9471)423.08nm — median 0.9411 (q25–q75 0.9319–0.9486)427.77nm — median 0.9425 (q25–q75 0.9334–0.9499)432.45nm — median 0.9437 (q25–q75 0.9345–0.9511)437.13nm — median 0.9442 (q25–q75 0.9351–0.9516)441.81nm — median 0.9438 (q25–q75 0.9346–0.9513)446.48nm — median 0.9421 (q25–q75 0.933–0.9502)451.15nm — median 0.9396 (q25–q75 0.9303–0.9476)455.82nm — median 0.9377 (q25–q75 0.9286–0.9461)460.49nm — median 0.9377 (q25–q75 0.9288–0.9462)465.15nm — median 0.938 (q25–q75 0.9294–0.9466)469.81nm — median 0.9381 (q25–q75 0.9297–0.9468)474.47nm — median 0.9374 (q25–q75 0.9292–0.9463)479.13nm — median 0.9361 (q25–q75 0.9278–0.9452)483.78nm — median 0.935 (q25–q75 0.9266–0.9443)488.43nm — median 0.9332 (q25–q75 0.9249–0.9429)493.08nm — median 0.9297 (q25–q75 0.921–0.9403)497.73nm — median 0.9234 (q25–q75 0.9136–0.9352)502.37nm — median 0.9134 (q25–q75 0.9018–0.9262)506.24nm — median 0.9003 (q25–q75 0.8856–0.9143)510.87nm — median 0.8768 (q25–q75 0.86–0.8928)515.51nm — median 0.8446 (q25–q75 0.8251–0.8633)520.14nm — median 0.8039 (q25–q75 0.7823–0.8295)524.77nm — median 0.7656 (q25–q75 0.74–0.7949)529.4nm — median 0.734 (q25–q75 0.7032–0.7656)534.02nm — median 0.7128 (q25–q75 0.679–0.7459)538.65nm — median 0.7003 (q25–q75 0.6653–0.7335)543.27nm — median 0.6924 (q25–q75 0.6566–0.7253)547.88nm — median 0.6852 (q25–q75 0.6479–0.7173)552.5nm — median 0.6802 (q25–q75 0.6423–0.7114)557.11nm — median 0.6826 (q25–q75 0.6445–0.714)561.72nm — median 0.6918 (q25–q75 0.6562–0.7245)566.32nm — median 0.7093 (q25–q75 0.6763–0.7409)570.93nm — median 0.7307 (q25–q75 0.7–0.7613)575.53nm — median 0.7515 (q25–q75 0.7243–0.781)580.13nm — median 0.7671 (q25–q75 0.743–0.7956)584.72nm — median 0.7802 (q25–q75 0.7567–0.8069)589.32nm — median 0.7904 (q25–q75 0.7655–0.8147)593.9nm — median 0.7962 (q25–q75 0.772–0.8197)598.49nm — median 0.7998 (q25–q75 0.776–0.823)603.08nm — median 0.8048 (q25–q75 0.7812–0.8273)607.66nm — median 0.8139 (q25–q75 0.7906–0.8357)612.24nm — median 0.8255 (q25–q75 0.8035–0.8463)616.82nm — median 0.8369 (q25–q75 0.8153–0.8568)621.39nm — median 0.8457 (q25–q75 0.8239–0.8643)625.96nm — median 0.8491 (q25–q75 0.8272–0.8674)630.53nm — median 0.8505 (q25–q75 0.8284–0.8681)635.1nm — median 0.855 (q25–q75 0.8332–0.8712)639.66nm — median 0.8664 (q25–q75 0.8465–0.8824)644.23nm — median 0.8821 (q25–q75 0.8641–0.8974)648.78nm — median 0.8946 (q25–q75 0.8789–0.9099)653.34nm — median 0.9021 (q25–q75 0.886–0.9159)657.89nm — median 0.9114 (q25–q75 0.8965–0.9237)662.44nm — median 0.9231 (q25–q75 0.9114–0.9346)666.99nm — median 0.9332 (q25–q75 0.9237–0.9434)671.54nm — median 0.9392 (q25–q75 0.9293–0.9483)676.08nm — median 0.9408 (q25–q75 0.9312–0.9496)680.62nm — median 0.9389 (q25–q75 0.9291–0.9477)685.16nm — median 0.9266 (q25–q75 0.917–0.9371)689.69nm — median 0.8867 (q25–q75 0.8695–0.9042)694.23nm — median 0.8059 (q25–q75 0.7802–0.8296)698.75nm — median 0.695 (q25–q75 0.6612–0.7266)703.28nm — median 0.5793 (q25–q75 0.5377–0.62)707.81nm — median 0.4764 (q25–q75 0.4266–0.5253)712.33nm — median 0.3828 (q25–q75 0.3281–0.4384)716.09nm — median 0.3123 (q25–q75 0.2492–0.3679)720.61nm — median 0.2356 (q25–q75 0.1645–0.294)725.12nm — median 0.1725 (q25–q75 0.09404–0.2249)729.64nm — median 0.118 (q25–q75 0.03489–0.1694)734.14nm — median 0.07525 (q25–q75 -0.006421–0.1305)738.65nm — median 0.04595 (q25–q75 -0.03595–0.1032)743.15nm — median 0.02548 (q25–q75 -0.06056–0.08642)747.65nm — median 0.01203 (q25–q75 -0.07571–0.07651)752.15nm — median 0.00258 (q25–q75 -0.08439–0.06743)756.65nm — median 0.0008833 (q25–q75 -0.08704–0.06377)761.14nm — median -0.003342 (q25–q75 -0.08762–0.06255)765.63nm — median -0.002291 (q25–q75 -0.08412–0.05798)770.12nm — median -0.0006572 (q25–q75 -0.08301–0.05554)774.61nm — median 0.002184 (q25–q75 -0.0771–0.05656)779.09nm — median 0.002526 (q25–q75 -0.07437–0.05974)783.57nm — median 0.004899 (q25–q75 -0.0726–0.06408)788.04nm — median 0.008407 (q25–q75 -0.07652–0.06628)792.52nm — median 0.009334 (q25–q75 -0.07273–0.06718)796.99nm — median 0.008415 (q25–q75 -0.07627–0.06638)801.46nm — median 0.009893 (q25–q75 -0.07426–0.06885)805.93nm — median 0.009051 (q25–q75 -0.07265–0.07058)810.39nm — median 0.009724 (q25–q75 -0.07392–0.07102)814.85nm — median 0.01391 (q25–q75 -0.06662–0.07299)819.31nm — median 0.01858 (q25–q75 -0.06237–0.07822)823.77nm — median 0.01853 (q25–q75 -0.06006–0.07743)828.22nm — median 0.02544 (q25–q75 -0.05398–0.08374)832.67nm — median 0.03274 (q25–q75 -0.04702–0.09235)837.12nm — median 0.03888 (q25–q75 -0.03902–0.0954)841.56nm — median 0.0471 (q25–q75 -0.03265–0.103)846.01nm — median 0.04812 (q25–q75 -0.02491–0.1058)850.45nm — median 0.05012 (q25–q75 -0.02005–0.1111)854.88nm — median 0.06432 (q25–q75 -0.01081–0.12)859.32nm — median 0.07341 (q25–q75 0.00464–0.1306)863.75nm — median 0.09251 (q25–q75 0.02433–0.1524)868.18nm — median 0.1154 (q25–q75 0.04183–0.1735)872.61nm — median 0.1293 (q25–q75 0.05226–0.1873)877.03nm — median 0.1413 (q25–q75 0.06622–0.2054)881.45nm — median 0.1579 (q25–q75 0.08928–0.2319)885.87nm — median 0.1848 (q25–q75 0.1169–0.252)890.29nm — median 0.2055 (q25–q75 0.1359–0.2743)894.7nm — median 0.2267 (q25–q75 0.1644–0.3005)899.11nm — median 0.2483 (q25–q75 0.1837–0.3263)903.52nm — median 0.2191 (q25–q75 0.1591–0.2927)907.93nm — median 0.2548 (q25–q75 0.1838–0.3253)912.33nm — median 0.2703 (q25–q75 0.2072–0.3448)916.73nm — median 0.3088 (q25–q75 0.2365–0.3755)920.4nm — median 0.3329 (q25–q75 0.2587–0.4045)924.79nm — median 0.3543 (q25–q75 0.2888–0.4256)929.18nm — median 0.3907 (q25–q75 0.3252–0.4653)933.58nm — median 0.4265 (q25–q75 0.36–0.4961)937.96nm — median 0.4532 (q25–q75 0.3736–0.5224)942.35nm — median 0.4747 (q25–q75 0.4036–0.5501)946.73nm — median 0.5002 (q25–q75 0.4273–0.5656)951.11nm — median 0.5498 (q25–q75 0.473–0.603)955.49nm — median 0.5689 (q25–q75 0.5011–0.6338)959.86nm — median 0.6007 (q25–q75 0.5327–0.6574)964.23nm — median 0.6183 (q25–q75 0.5595–0.6769)968.6nm — median 0.6412 (q25–q75 0.5835–0.692)972.97nm — median 0.6578 (q25–q75 0.6127–0.7097)977.33nm — median 0.6767 (q25–q75 0.626–0.7265)981.69nm — median 0.6936 (q25–q75 0.654–0.7402)986.05nm — median 0.7102 (q25–q75 0.6633–0.7492)990.41nm — median 0.7196 (q25–q75 0.6714–0.7573)994.76nm — median 0.7274 (q25–q75 0.6907–0.7631)999.11nm — median 0.7303 (q25–q75 0.6918–0.7634)1003.5nm — median 0.7301 (q25–q75 0.7006–0.764)1007.8nm — median 0.7363 (q25–q75 0.7022–0.766)1012.1nm — median 0.7372 (q25–q75 0.7081–0.7688)1016.5nm — median 0.7359 (q25–q75 0.7055–0.7651)1020.8nm — median 0.7336 (q25–q75 0.7076–0.7639)

Sampling

Wavelengths832
Axis range394.9–1021 nm
Mean spacing0.753 nm
Gridirregular
Observations216

Signal & quality

Value range-0.74 – 1.03
Mean range-0.0175 – 0.978
Mean level0.5449
Area343.9
PTP0.9955
Noise RMS0.00055937
SNR1e+03
SNR dB6e+01 dB
Dynamic range0.995
Smoothness0.0182
Saturated0.0%
X-outliers117

Integrity & artefacts

NaN ratio0.00%
Inf count0
Zero ratio0.00%
Spike count40,627
Spike rate22.66%
Jump count4,884
Jump rate2.72%
Clip fraction0.00%

Shape & reference

Baseline slope-0.72914
Curvature RMS0.017974
D1 RMS0.011535
RMS to mean0.054391
RMS p950.17189
SAM to mean0.074573
SAM p950.24289
Affine offset p950.29277
Affine gain p95 Δ0.30381
Affine residual p950.037936
Xcorr lag p950

Outliers & repeatability

PCA Q p95/median2.4
Hotelling T2 p95/median4.2
Mahalanobis H p95/median2.1
Repeat groups0

Dimensionality (PCA)

Effective rank1.6
PCs → 95% var3
PCs → 99% var4
Top-10 cum. var99.3%
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_reflectance0.544881.00fortValeur 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_curve343.881.00fortValeur 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_peak0.995480.00faibleVariabilité forteSaturationmax(mean_spectrum) - min(mean_spectrum)alert increases when dynamic range is abnormally flat
Amplitude globaleVarianceamplitude.variance0.138290.00faibleNormal ou hétérogèneMauvais contactvar(X finite)alert increases when variance/dynamic range is abnormally flat
BruitNoise RMSnoise.noise_rms0.000559370.01faibleStableLampe, détecteurmedian MAD(second derivative) * 1.4826 / sqrt(6)alert = noise_rms / signal_scale, saturated at 5%
BruitSNRnoise.snr1013.60.00faibleBon signalAcquisitionmean(abs(X)) / noise_rmsalert decreases with SNR dB; >=40 dB is treated as low alert
BruitBandwise SNRnoise.bandwise_snr_min0.118461.00fortZone problématiqueDétecteurmin(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_count40,6271.00fortArtefactsCosmic rays, splicecount robust outliers in second derivativealert follows spike_rate, saturated at 1%
Artefacts locauxSpike rateartefacts.spike_rate22.7%1.00fortSpectre suspectInterpolationspike_count / (n_samples * (n_features - 2))alert = min(1, spike_rate / 0.01)
Artefacts locauxJump countartefacts.jump_count4,8841.00fortRaccord détecteurSplicecount robust outliers in first derivativealert follows jump_rate, saturated at 1%
Artefacts locauxJump rateartefacts.jump_rate2.72%1.00fortProblème spectralCalibrationjump_count / (n_samples * (n_features - 1))alert = min(1, jump_rate / 0.01)
Artefacts locauxClip fractionartefacts.clip_fraction0.00111%0.00faibleNormalDétecteur saturéfraction of finite cells equal to repeated min/max extremaalert = min(1, clip_fraction / 0.01)
Forme spectraleBaseline slopeshape.baseline_slope-0.729141.00fortDériveÉclairementlinear slope of mean_spectrum over normalized axisalert = abs(slope / signal_scale), saturated at 0.5
Forme spectraleCurvature RMSshape.curvature_rms0.0179741.00fortForme inhabituelleFond, splicemedian RMS(second derivative per spectrum)alert = curvature_rms / signal_scale, saturated at 1%
Forme spectraleD1 RMSshape.d1_rms0.0115350.23faiblePlatBiologie ou artefactmedian RMS(first derivative per spectrum)alert = d1_rms / signal_scale, saturated at 5%
Outliers multivariésPCA Q (SPE)outliers.pca_q_ratio2.36090.30faibleConformeArtefact, mélangep95(Q/SPE residual) / median(Q/SPE residual)alert = min(1, pca_q_ratio / 8)
Outliers multivariésHotelling T²outliers.hotelling_t2_ratio4.24250.53moyenExtrême mais cohérentVariabilité naturellep95(Hotelling T2) / median(Hotelling T2)alert = min(1, hotelling_t2_ratio / 8)
Outliers multivariésMahalanobis Houtliers.mahalanobis_h_ratio2.05270.51moyenOutlier globalDomaine différentp95(sqrt(T2)) / median(sqrt(T2))alert = min(1, mahalanobis_h_ratio / 4)
Comparaison à référenceRMS to mean spectrumreference.rms_to_mean_spectrum_p950.171890.69moyenSpectre différentDomain 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_p950.242890.69moyenForme différenteFond, 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_density2.54671.00fortSous-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_p954.02381.00fortSpectre 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_p950.590761.00fortSpectre atypiqueDiverses causesp95 IsolationForest anomaly score on PCA scoresalert follows structure complexity; raw score is implementation-dependent
X_absorbance PCA score plot-15-10-50510-2-101PC1 -0.3462 · PC2 -0.8517PC1 -5.465 · PC2 -1.412PC1 -2.454 · PC2 -1.03PC1 -3.844 · PC2 -1.253PC1 0.654 · PC2 -0.7645PC1 -7.949 · PC2 -1.522PC1 -0.1106 · PC2 0.4848PC1 -3.802 · PC2 0.6319PC1 -1.588 · PC2 0.795PC1 -3.659 · PC2 0.4546PC1 -0.5743 · PC2 0.5275PC1 -12.34 · PC2 0.3295PC1 3.789 · PC2 -0.3445PC1 1.67 · PC2 -0.6606PC1 0.7591 · PC2 -0.7161PC1 0.2253 · PC2 -0.7829PC1 0.7671 · PC2 -0.7186PC1 4.475 · PC2 -0.3446PC1 0.04032 · PC2 -0.7522PC1 1.731 · PC2 -0.5602PC1 1.079 · PC2 -0.536PC1 0.04554 · PC2 -0.6729PC1 1.931 · PC2 -0.4745PC1 1.597 · PC2 -0.6745PC1 -2.041 · PC2 -0.8869PC1 0.6852 · PC2 -0.6732PC1 1.131 · PC2 -0.6264PC1 1.202 · PC2 -0.6325PC1 -0.3758 · PC2 -0.7347PC1 2.028 · PC2 -0.4831PC1 2.08 · PC2 -0.5173PC1 3.344 · PC2 -0.3845PC1 1.155 · PC2 -0.6969PC1 0.4464 · PC2 -0.7634PC1 -0.5902 · PC2 -0.8115PC1 3.329 · PC2 -0.4352PC1 1.428 · PC2 -0.5843PC1 -0.7084 · PC2 -0.8454PC1 0.9887 · PC2 -0.6114PC1 -1.369 · PC2 -0.8087PC1 5.596 · PC2 -0.1675PC1 2.414 · PC2 -0.4574PC1 3.395 · PC2 0.7968PC1 1.18 · PC2 0.6393PC1 0.0801 · PC2 0.6137PC1 -0.4364 · PC2 0.6419PC1 0.2053 · PC2 0.5749PC1 4.208 · PC2 0.7557PC1 -0.539 · PC2 0.637PC1 1.316 · PC2 0.8052PC1 0.2553 · PC2 0.6611PC1 -0.5126 · PC2 0.7928PC1 1.101 · PC2 0.9013PC1 -0.01652 · PC2 0.5227PC1 -2.686 · PC2 0.7095PC1 0.1649 · PC2 0.7626PC1 0.3586 · PC2 0.7128PC1 -6.091 · PC2 0.5811PC1 -0.9897 · PC2 0.6978PC1 1.386 · PC2 0.7314PC1 1.557 · PC2 0.6709PC1 2.936 · PC2 0.748PC1 -0.06941 · PC2 0.4465PC1 -0.393 · PC2 0.5956PC1 -1.847 · PC2 0.5278PC1 2.508 · PC2 0.6492PC1 1.021 · PC2 0.7794PC1 -1.495 · PC2 0.6093PC1 0.3908 · PC2 0.6339PC1 -2.528 · PC2 0.6593PC1 4.992 · PC2 0.6384PC1 1.797 · PC2 0.7747PC1 0.7945 · PC2 -0.7189PC1 1.114 · PC2 -0.6957PC1 -1.138 · PC2 -0.8881PC1 -0.2196 · PC2 -0.7745PC1 1.706 · PC2 -0.508PC1 -1.464 · PC2 -0.8265PC1 -0.1675 · PC2 -0.7533PC1 -0.5848 · PC2 -0.8465PC1 1.494 · PC2 -0.458PC1 -1.65 · PC2 -0.8302PC1 0.3883 · PC2 -0.7003PC1 1.706 · PC2 -0.4661PC1 -0.779 · PC2 -0.8823PC1 -0.04246 · PC2 -0.795PC1 0.122 · PC2 -0.6785PC1 1.323 · PC2 -0.4519PC1 -1.16 · PC2 -0.8762PC1 0.4767 · PC2 -0.6759PC1 1.625 · PC2 -0.5689PC1 1.593 · PC2 -0.5139PC1 1.808 · PC2 -0.4515PC1 -2.441 · PC2 -0.8348PC1 2.895 · PC2 -0.3873PC1 4.859 · PC2 -0.2534PC1 1.367 · PC2 -0.661PC1 -3.135 · PC2 -1.068PC1 -1.507 · PC2 -0.8625PC1 1.638 · PC2 -0.5852PC1 0.6721 · PC2 -0.6162PC1 0.8171 · PC2 -0.5358PC1 -0.5765 · PC2 -0.7298PC1 -0.5492 · PC2 -0.7163PC1 -0.8492 · PC2 -0.7973PC1 -0.9754 · PC2 -0.8721PC1 3.332 · PC2 -0.4221PC1 2.386 · PC2 -0.4847PC1 0.4656 · PC2 0.7594PC1 0.4839 · PC2 0.5829PC1 -1.875 · PC2 0.5944PC1 -0.9063 · PC2 0.5777PC1 0.9655 · PC2 0.6582PC1 -2.005 · PC2 0.6786PC1 -0.5049 · PC2 0.7058PC1 -1.09 · PC2 0.5464PC1 0.5349 · PC2 0.8232PC1 -2.667 · PC2 0.5735PC1 -0.1389 · PC2 0.6332PC1 1.196 · PC2 0.9571PC1 -1.342 · PC2 0.5608PC1 -1.124 · PC2 0.5764PC1 -0.5908 · PC2 0.6417PC1 0.7358 · PC2 0.7237PC1 -2.085 · PC2 0.5652PC1 -0.05374 · PC2 0.6564PC1 0.9576 · PC2 0.6768PC1 0.5224 · PC2 0.7207PC1 1.311 · PC2 0.8409PC1 -3.435 · PC2 0.6621PC1 2.034 · PC2 0.7904PC1 4.432 · PC2 0.8253PC1 0.9461 · PC2 0.7028PC1 -4.186 · PC2 0.4565PC1 -2.253 · PC2 0.7178PC1 0.7912 · PC2 0.6299PC1 -0.05369 · PC2 0.6455PC1 0.2292 · PC2 0.7435PC1 -1.494 · PC2 0.6297PC1 -1.314 · PC2 0.6887PC1 -12.51 · PC2 0.3352PC1 -5.831 · PC2 0.4781PC1 2.458 · PC2 0.6624PC1 1.35 · PC2 0.6547PC1 0.6989 · PC2 -0.7398PC1 -2.59 · PC2 -1.021PC1 3.045 · PC2 -0.3526PC1 -0.5058 · PC2 -0.7592PC1 -0.857 · PC2 -0.7362PC1 1.331 · PC2 -0.5187PC1 -1.28 · PC2 -0.8703PC1 -2.708 · PC2 -0.8704PC1 -1.38 · PC2 -0.7798PC1 0.364 · PC2 -0.688PC1 2.071 · PC2 -0.4898PC1 0.9585 · PC2 -0.5515PC1 -0.04071 · PC2 -0.6928PC1 1.601 · PC2 -0.5238PC1 1.148 · PC2 -0.347PC1 1.004 · PC2 -0.5816PC1 3.111 · PC2 -0.3786PC1 -1.255 · PC2 -0.6611PC1 3.516 · PC2 -0.3394PC1 -2.028 · PC2 -0.8002PC1 -1.007 · PC2 -0.7396PC1 6.146 · PC2 -0.01452PC1 1.153 · PC2 -0.4804PC1 1.014 · PC2 -0.4886PC1 0.03745 · PC2 -0.7255PC1 -0.9196 · PC2 -0.8478PC1 0.3288 · PC2 -0.3521PC1 0.5171 · PC2 -0.543PC1 -0.2282 · PC2 -0.6984PC1 -2.268 · PC2 -0.7657PC1 -1.638 · PC2 -0.7943PC1 1.084 · PC2 -0.554PC1 0.8806 · PC2 -0.5577PC1 -0.2203 · PC2 -0.5991PC1 2.551 · PC2 -0.3808PC1 2.021 · PC2 -0.4571PC1 -0.1656 · PC2 0.5281PC1 -6.657 · PC2 0.5996PC1 2.379 · PC2 0.7212PC1 -0.6047 · PC2 0.8168PC1 -2.029 · PC2 0.5885PC1 1.35 · PC2 0.7769PC1 -2.429 · PC2 0.5397PC1 -4.282 · PC2 0.7868PC1 -2.072 · PC2 0.6885PC1 0.1101 · PC2 0.7192PC1 3.17 · PC2 0.8145PC1 0.06212 · PC2 0.6905PC1 0.5123 · PC2 0.6182PC1 0.9076 · PC2 0.6712PC1 0.5273 · PC2 0.8918PC1 0.03941 · PC2 0.5721PC1 2.522 · PC2 0.6579PC1 -2.158 · PC2 0.6501PC1 2.376 · PC2 0.6255PC1 -3.061 · PC2 0.5695PC1 -1.764 · PC2 0.5874PC1 5.504 · PC2 0.8285PC1 0.1901 · PC2 0.6894PC1 0.3571 · PC2 0.732PC1 -0.9726 · PC2 0.5557PC1 -2.196 · PC2 0.4686PC1 -0.335 · PC2 0.8219PC1 0.3005 · PC2 0.7264PC1 -1.252 · PC2 0.4753PC1 -3.024 · PC2 0.6372PC1 -2.434 · PC2 0.5753PC1 0.0283 · PC2 0.6121PC1 0.0515 · PC2 0.707PC1 -1.12 · PC2 0.6702PC1 2.011 · PC2 0.6358PC1 1.05 · PC2 0.6593PC1 (88.0%)PC2 (6.9%)216 scores
PCA explained variance0%25%50%75%100%PC1: 88.0% (cumulative 88.0%)1PC2: 6.9% (cumulative 94.9%)2PC3: 3.9% (cumulative 98.7%)3PC4: 0.3% (cumulative 99.0%)4PC5: 0.1% (cumulative 99.2%)5PC6: 0.1% (cumulative 99.2%)6PC7: 0.0% (cumulative 99.3%)7PC8: 0.0% (cumulative 99.3%)8PC9: 0.0% (cumulative 99.3%)9PC10: 0.0% (cumulative 99.3%)10cumulative explained variancePC variancecumulativeprincipal component · cumulative (dashed)

reflectance

X_reflectance · NIR
reflectance spectra-0.20.00.20.40.62004006008001,0001,200q05-q95 envelopeq25-q75 envelopemedian spectrummedianq25–q75q05–q95wavelength / nm394.92nm — median 0.01163 (q25–q75 0.007362–0.01563)399.62nm — median 0.05406 (q25–q75 0.04717–0.06216)404.31nm — median 0.05337 (q25–q75 0.04636–0.06142)409.01nm — median 0.05257 (q25–q75 0.0455–0.0601)413.7nm — median 0.05153 (q25–q75 0.04499–0.05975)418.39nm — median 0.05113 (q25–q75 0.04458–0.05958)423.08nm — median 0.05081 (q25–q75 0.04427–0.05907)427.77nm — median 0.05041 (q25–q75 0.0438–0.05892)432.45nm — median 0.05021 (q25–q75 0.04367–0.05868)437.13nm — median 0.04981 (q25–q75 0.04338–0.05828)441.81nm — median 0.04976 (q25–q75 0.04329–0.05811)446.48nm — median 0.05017 (q25–q75 0.04336–0.05841)451.15nm — median 0.05081 (q25–q75 0.04351–0.05879)455.82nm — median 0.05097 (q25–q75 0.04382–0.05914)460.49nm — median 0.05104 (q25–q75 0.04361–0.05895)465.15nm — median 0.05095 (q25–q75 0.04334–0.0587)469.81nm — median 0.0508 (q25–q75 0.04324–0.05857)474.47nm — median 0.0508 (q25–q75 0.04331–0.05843)479.13nm — median 0.05099 (q25–q75 0.04344–0.05861)483.78nm — median 0.05103 (q25–q75 0.04359–0.05873)488.43nm — median 0.05131 (q25–q75 0.04376–0.05914)493.08nm — median 0.05192 (q25–q75 0.04427–0.05981)497.73nm — median 0.05308 (q25–q75 0.0453–0.06103)502.37nm — median 0.05544 (q25–q75 0.04749–0.06418)506.24nm — median 0.0584 (q25–q75 0.0502–0.06807)510.87nm — median 0.06463 (q25–q75 0.05663–0.07646)515.51nm — median 0.07485 (q25–q75 0.06678–0.08926)520.14nm — median 0.08972 (q25–q75 0.08024–0.1058)524.77nm — median 0.1075 (q25–q75 0.09538–0.125)529.4nm — median 0.1235 (q25–q75 0.1089–0.141)534.02nm — median 0.1334 (q25–q75 0.1178–0.152)538.65nm — median 0.1391 (q25–q75 0.1231–0.1585)543.27nm — median 0.1428 (q25–q75 0.1262–0.1624)547.88nm — median 0.1463 (q25–q75 0.129–0.1661)552.5nm — median 0.1487 (q25–q75 0.131–0.1683)557.11nm — median 0.1469 (q25–q75 0.1299–0.167)561.72nm — median 0.1411 (q25–q75 0.125–0.1608)566.32nm — median 0.1319 (q25–q75 0.1172–0.1515)570.93nm — median 0.1195 (q25–q75 0.1073–0.1391)575.53nm — median 0.1089 (q25–q75 0.09824–0.1275)580.13nm — median 0.1011 (q25–q75 0.09065–0.1193)584.72nm — median 0.09535 (q25–q75 0.08541–0.1131)589.32nm — median 0.09136 (q25–q75 0.08202–0.1091)593.9nm — median 0.08877 (q25–q75 0.07995–0.1063)598.49nm — median 0.08742 (q25–q75 0.07877–0.1048)603.08nm — median 0.08554 (q25–q75 0.07701–0.1028)607.66nm — median 0.08192 (q25–q75 0.07361–0.09892)612.24nm — median 0.07754 (q25–q75 0.06907–0.09358)616.82nm — median 0.07292 (q25–q75 0.06515–0.08876)621.39nm — median 0.06985 (q25–q75 0.06211–0.08525)625.96nm — median 0.06847 (q25–q75 0.06093–0.08337)630.53nm — median 0.06796 (q25–q75 0.06044–0.08282)635.1nm — median 0.06673 (q25–q75 0.0592–0.08088)639.66nm — median 0.0635 (q25–q75 0.05623–0.0763)644.23nm — median 0.05895 (q25–q75 0.05157–0.07054)648.78nm — median 0.05532 (q25–q75 0.04806–0.06595)653.34nm — median 0.05382 (q25–q75 0.04657–0.06345)657.89nm — median 0.05192 (q25–q75 0.04421–0.0603)662.44nm — median 0.04956 (q25–q75 0.04191–0.0571)666.99nm — median 0.04762 (q25–q75 0.04028–0.05417)671.54nm — median 0.0469 (q25–q75 0.03979–0.05354)676.08nm — median 0.04692 (q25–q75 0.0397–0.05366)680.62nm — median 0.04773 (q25–q75 0.0402–0.05445)685.16nm — median 0.05058 (q25–q75 0.04215–0.0576)689.69nm — median 0.05878 (q25–q75 0.04994–0.0695)694.23nm — median 0.08356 (q25–q75 0.0732–0.1004)698.75nm — median 0.1271 (q25–q75 0.1147–0.1508)703.28nm — median 0.1809 (q25–q75 0.1625–0.2085)707.81nm — median 0.2321 (q25–q75 0.2107–0.2658)712.33nm — median 0.2811 (q25–q75 0.2556–0.3178)716.09nm — median 0.3207 (q25–q75 0.2923–0.3574)720.61nm — median 0.3627 (q25–q75 0.3302–0.4)725.12nm — median 0.3981 (q25–q75 0.364–0.4348)729.64nm — median 0.4262 (q25–q75 0.3908–0.4632)734.14nm — median 0.4488 (q25–q75 0.4136–0.482)738.65nm — median 0.4649 (q25–q75 0.4293–0.499)743.15nm — median 0.4776 (q25–q75 0.4402–0.5109)747.65nm — median 0.4823 (q25–q75 0.4456–0.5195)752.15nm — median 0.4862 (q25–q75 0.4492–0.524)756.65nm — median 0.4887 (q25–q75 0.4514–0.5256)761.14nm — median 0.4879 (q25–q75 0.4534–0.5263)765.63nm — median 0.4877 (q25–q75 0.4537–0.5251)770.12nm — median 0.4869 (q25–q75 0.4524–0.5239)774.61nm — median 0.4853 (q25–q75 0.4511–0.5224)779.09nm — median 0.4826 (q25–q75 0.4497–0.5211)783.57nm — median 0.481 (q25–q75 0.4495–0.5197)788.04nm — median 0.4808 (q25–q75 0.4478–0.5214)792.52nm — median 0.4794 (q25–q75 0.4479–0.5206)796.99nm — median 0.4782 (q25–q75 0.4459–0.5189)801.46nm — median 0.4774 (q25–q75 0.4451–0.518)805.93nm — median 0.4779 (q25–q75 0.4454–0.5187)810.39nm — median 0.4766 (q25–q75 0.4441–0.5166)814.85nm — median 0.4743 (q25–q75 0.443–0.5145)819.31nm — median 0.4744 (q25–q75 0.4421–0.5145)823.77nm — median 0.4736 (q25–q75 0.4418–0.513)828.22nm — median 0.4699 (q25–q75 0.4392–0.5116)832.67nm — median 0.4694 (q25–q75 0.437–0.5085)837.12nm — median 0.4662 (q25–q75 0.4347–0.5058)841.56nm — median 0.4625 (q25–q75 0.432–0.5036)846.01nm — median 0.4572 (q25–q75 0.4274–0.4985)850.45nm — median 0.452 (q25–q75 0.4215–0.4934)854.88nm — median 0.446 (q25–q75 0.4173–0.487)859.32nm — median 0.4436 (q25–q75 0.4142–0.482)863.75nm — median 0.4367 (q25–q75 0.4075–0.4757)868.18nm — median 0.4286 (q25–q75 0.3972–0.4663)872.61nm — median 0.4204 (q25–q75 0.3886–0.458)877.03nm — median 0.409 (q25–q75 0.3778–0.4476)881.45nm — median 0.3994 (q25–q75 0.3702–0.437)885.87nm — median 0.3948 (q25–q75 0.3604–0.4316)890.29nm — median 0.3856 (q25–q75 0.3513–0.4227)894.7nm — median 0.3714 (q25–q75 0.3349–0.4093)899.11nm — median 0.3563 (q25–q75 0.3226–0.3902)903.52nm — median 0.3775 (q25–q75 0.3446–0.4148)907.93nm — median 0.3629 (q25–q75 0.3278–0.3999)912.33nm — median 0.3532 (q25–q75 0.316–0.3885)916.73nm — median 0.3385 (q25–q75 0.3032–0.367)920.4nm — median 0.3273 (q25–q75 0.2917–0.3629)924.79nm — median 0.3121 (q25–q75 0.278–0.3445)929.18nm — median 0.296 (q25–q75 0.2573–0.3294)933.58nm — median 0.2789 (q25–q75 0.2446–0.3106)937.96nm — median 0.2688 (q25–q75 0.2292–0.2965)942.35nm — median 0.2546 (q25–q75 0.2226–0.2892)946.73nm — median 0.2433 (q25–q75 0.212–0.2764)951.11nm — median 0.2242 (q25–q75 0.194–0.2519)955.49nm — median 0.2086 (q25–q75 0.181–0.2404)959.86nm — median 0.1971 (q25–q75 0.1689–0.2245)964.23nm — median 0.1861 (q25–q75 0.1578–0.2117)968.6nm — median 0.1756 (q25–q75 0.1529–0.2017)972.97nm — median 0.1654 (q25–q75 0.1438–0.1921)977.33nm — median 0.1592 (q25–q75 0.1374–0.1827)981.69nm — median 0.1527 (q25–q75 0.13–0.1744)986.05nm — median 0.1505 (q25–q75 0.1289–0.1651)990.41nm — median 0.1459 (q25–q75 0.126–0.1642)994.76nm — median 0.1429 (q25–q75 0.1264–0.1612)999.11nm — median 0.1418 (q25–q75 0.1265–0.1615)1003.5nm — median 0.1426 (q25–q75 0.1252–0.1602)1007.8nm — median 0.1425 (q25–q75 0.1268–0.16)1012.1nm — median 0.1427 (q25–q75 0.1268–0.1582)1016.5nm — median 0.1429 (q25–q75 0.1283–0.1607)1020.8nm — median 0.1458 (q25–q75 0.1295–0.1617)

Sampling

Wavelengths832
Axis range394.9–1021 nm
Mean spacing0.753 nm
Gridirregular
Observations216

Signal & quality

Value range-0.0178 – 0.891
Mean range0.0115 – 0.493
Mean level0.2238
Area138.8
PTP0.4811
Noise RMS0.00029368
SNR7.6e+02
SNR dB6e+01 dB
Dynamic range0.481
Smoothness0.009841
Saturated0.0%
X-outliers115

Integrity & artefacts

NaN ratio0.00%
Inf count0
Zero ratio0.00%
Spike count40,139
Spike rate22.39%
Jump count6,871
Jump rate3.83%
Clip fraction0.00%

Shape & reference

Baseline slope0.34019
Curvature RMS0.0096796
D1 RMS0.005992
RMS to mean0.028078
RMS p950.083664
SAM to mean0.04982
SAM p950.078081
Affine offset p950.026494
Affine gain p95 Δ0.32029
Affine residual p950.019147
Xcorr lag p950

Outliers & repeatability

PCA Q p95/median3.2
Hotelling T2 p95/median3.7
Mahalanobis H p95/median1.9
Repeat groups0

Dimensionality (PCA)

Effective rank1.6
PCs → 95% var3
PCs → 99% var5
Top-10 cum. var99.2%
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_ratio0.000556%0.00faibleNormalExport, saturationcount(X == 0) / count(finite X)alert = min(1, zero_ratio / 0.05)
Amplitude globaleMean reflectanceamplitude.mean_reflectance0.223831.00fortValeur 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_curve138.781.00fortValeur 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_peak0.481110.00faibleVariabilité forteSaturationmax(mean_spectrum) - min(mean_spectrum)alert increases when dynamic range is abnormally flat
Amplitude globaleVarianceamplitude.variance0.0306530.00faibleNormal ou hétérogèneMauvais contactvar(X finite)alert increases when variance/dynamic range is abnormally flat
BruitNoise RMSnoise.noise_rms0.000293680.01faibleStableLampe, détecteurmedian MAD(second derivative) * 1.4826 / sqrt(6)alert = noise_rms / signal_scale, saturated at 5%
BruitSNRnoise.snr762.170.00faibleBon signalAcquisitionmean(abs(X)) / noise_rmsalert decreases with SNR dB; >=40 dB is treated as low alert
BruitBandwise SNRnoise.bandwise_snr_min1.88980.84fortZone problématiqueDétecteurmin(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_count40,1391.00fortArtefactsCosmic rays, splicecount robust outliers in second derivativealert follows spike_rate, saturated at 1%
Artefacts locauxSpike rateartefacts.spike_rate22.4%1.00fortSpectre suspectInterpolationspike_count / (n_samples * (n_features - 2))alert = min(1, spike_rate / 0.01)
Artefacts locauxJump countartefacts.jump_count6,8711.00fortRaccord détecteurSplicecount robust outliers in first derivativealert follows jump_rate, saturated at 1%
Artefacts locauxJump rateartefacts.jump_rate3.83%1.00fortProblème spectralCalibrationjump_count / (n_samples * (n_features - 1))alert = min(1, jump_rate / 0.01)
Artefacts locauxClip fractionartefacts.clip_fraction0.00111%0.00faibleNormalDétecteur saturéfraction of finite cells equal to repeated min/max extremaalert = min(1, clip_fraction / 0.01)
Forme spectraleBaseline slopeshape.baseline_slope0.340191.00fortDériveÉclairementlinear slope of mean_spectrum over normalized axisalert = abs(slope / signal_scale), saturated at 0.5
Forme spectraleCurvature RMSshape.curvature_rms0.00967961.00fortForme inhabituelleFond, splicemedian RMS(second derivative per spectrum)alert = curvature_rms / signal_scale, saturated at 1%
Forme spectraleD1 RMSshape.d1_rms0.0059920.25faiblePlatBiologie ou artefactmedian RMS(first derivative per spectrum)alert = d1_rms / signal_scale, saturated at 5%
Outliers multivariésPCA Q (SPE)outliers.pca_q_ratio3.23650.40faibleConformeArtefact, mélangep95(Q/SPE residual) / median(Q/SPE residual)alert = min(1, pca_q_ratio / 8)
Outliers multivariésHotelling T²outliers.hotelling_t2_ratio3.65810.46moyenExtrême mais cohérentVariabilité naturellep95(Hotelling T2) / median(Hotelling T2)alert = min(1, hotelling_t2_ratio / 8)
Outliers multivariésMahalanobis Houtliers.mahalanobis_h_ratio1.91160.48moyenOutlier globalDomaine différentp95(sqrt(T2)) / median(sqrt(T2))alert = min(1, mahalanobis_h_ratio / 4)
Comparaison à référenceRMS to mean spectrumreference.rms_to_mean_spectrum_p950.0836640.70moyenSpectre différentDomain 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_p950.0780810.22faibleSimilaireFond, 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_density4.83771.00fortSous-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_p953.25061.00fortSpectre 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_p950.56881.00fortSpectre atypiqueDiverses causesp95 IsolationForest anomaly score on PCA scoresalert follows structure complexity; raw score is implementation-dependent
X_reflectance PCA score plot-7.5-5.0-2.50.02.55.0-1.0-0.50.00.5PC1 -0.1423 · PC2 -0.3773PC1 -1.935 · PC2 -0.5196PC1 -0.4681 · PC2 -0.3782PC1 -1.211 · PC2 -0.486PC1 0.7566 · PC2 -0.3266PC1 -3.56 · PC2 -0.6298PC1 0.2147 · PC2 0.2066PC1 -0.6294 · PC2 0.3718PC1 0.02749 · PC2 0.3831PC1 -0.9148 · PC2 0.2122PC1 0.3746 · PC2 0.2551PC1 -5.515 · PC2 0.1914PC1 1.89 · PC2 -0.1369PC1 1.475 · PC2 -0.2517PC1 0.6155 · PC2 -0.3218PC1 -0.1548 · PC2 -0.3745PC1 0.4621 · PC2 -0.3379PC1 2.652 · PC2 -0.1734PC1 -0.1872 · PC2 -0.3637PC1 0.4681 · PC2 -0.2826PC1 1.124 · PC2 -0.1706PC1 -0.1726 · PC2 -0.3371PC1 0.8062 · PC2 -0.2311PC1 -0.06749 · PC2 -0.4211PC1 -1.981 · PC2 -0.4983PC1 0.5506 · PC2 -0.2784PC1 0.7951 · PC2 -0.2739PC1 0.6858 · PC2 -0.2832PC1 0.05105 · PC2 -0.3244PC1 1.293 · PC2 -0.1962PC1 0.8849 · PC2 -0.2546PC1 1.418 · PC2 -0.2335PC1 0.319 · PC2 -0.3356PC1 0.1368 · PC2 -0.3574PC1 0.6965 · PC2 -0.2705PC1 1.139 · PC2 -0.2899PC1 0.7995 · PC2 -0.2659PC1 -0.2071 · PC2 -0.4435PC1 1.404 · PC2 -0.2052PC1 -0.0852 · PC2 -0.2574PC1 2.953 · PC2 -0.1393PC1 0.8566 · PC2 -0.2692PC1 1.811 · PC2 0.3991PC1 1.36 · PC2 0.3061PC1 0.4066 · PC2 0.288PC1 -0.3335 · PC2 0.3431PC1 0.1951 · PC2 0.2238PC1 2.616 · PC2 0.275PC1 -0.4387 · PC2 0.269PC1 0.3532 · PC2 0.3835PC1 0.8335 · PC2 0.3049PC1 -0.3081 · PC2 0.3929PC1 0.4813 · PC2 0.418PC1 -0.7887 · PC2 0.2309PC1 -2.014 · PC2 0.3717PC1 0.4811 · PC2 0.4048PC1 0.5303 · PC2 0.3323PC1 -2.662 · PC2 0.3227PC1 -0.1057 · PC2 0.3274PC1 1.103 · PC2 0.3728PC1 0.7082 · PC2 0.3099PC1 1.235 · PC2 0.2825PC1 -0.2158 · PC2 0.2139PC1 -0.1581 · PC2 0.3023PC1 0.1877 · PC2 0.25PC1 0.9568 · PC2 0.2718PC1 0.7452 · PC2 0.3839PC1 -0.4506 · PC2 0.2687PC1 1.255 · PC2 0.327PC1 -0.4978 · PC2 0.3702PC1 2.684 · PC2 0.1991PC1 0.6151 · PC2 0.3396PC1 0.3916 · PC2 -0.3058PC1 1.002 · PC2 -0.2348PC1 -0.4275 · PC2 -0.3541PC1 0.3245 · PC2 -0.2978PC1 0.7454 · PC2 -0.2201PC1 -0.8691 · PC2 -0.3663PC1 0.3847 · PC2 -0.2932PC1 -0.5382 · PC2 -0.4261PC1 0.8354 · PC2 -0.148PC1 -0.9134 · PC2 -0.3556PC1 0.2711 · PC2 -0.3156PC1 0.9718 · PC2 -0.1656PC1 -1.142 · PC2 -0.4776PC1 -0.008404 · PC2 -0.3548PC1 -0.7869 · PC2 -0.4057PC1 -0.1375 · PC2 -0.2665PC1 -1.259 · PC2 -0.4787PC1 0.1695 · PC2 -0.36PC1 0.137 · PC2 -0.3337PC1 1.174 · PC2 -0.1858PC1 0.8933 · PC2 -0.192PC1 -1.202 · PC2 -0.3184PC1 0.8349 · PC2 -0.2521PC1 1.926 · PC2 -0.1872PC1 -0.03889 · PC2 -0.3639PC1 -1.493 · PC2 -0.441PC1 -1.224 · PC2 -0.4197PC1 0.5387 · PC2 -0.2991PC1 -0.2369 · PC2 -0.3216PC1 0.3832 · PC2 -0.2298PC1 -0.2173 · PC2 -0.3206PC1 -0.5341 · PC2 -0.3291PC1 -1.055 · PC2 -0.4535PC1 -0.9582 · PC2 -0.4475PC1 1.172 · PC2 -0.2516PC1 0.9743 · PC2 -0.2591PC1 0.3382 · PC2 0.3935PC1 0.8154 · PC2 0.32PC1 -0.736 · PC2 0.311PC1 0.08064 · PC2 0.2907PC1 0.4905 · PC2 0.3291PC1 -1.034 · PC2 0.3384PC1 0.33 · PC2 0.3507PC1 -0.6752 · PC2 0.2378PC1 0.5563 · PC2 0.4402PC1 -1.255 · PC2 0.307PC1 0.09509 · PC2 0.2812PC1 0.8324 · PC2 0.4556PC1 -1.359 · PC2 0.2526PC1 -0.4295 · PC2 0.2938PC1 -1.097 · PC2 0.2805PC1 -0.4206 · PC2 0.331PC1 -1.69 · PC2 0.2184PC1 0.01779 · PC2 0.2713PC1 -0.132 · PC2 0.2782PC1 0.802 · PC2 0.3555PC1 0.7747 · PC2 0.4026PC1 -1.525 · PC2 0.3786PC1 0.4719 · PC2 0.3398PC1 1.755 · PC2 0.3246PC1 -0.1721 · PC2 0.3155PC1 -1.911 · PC2 0.2655PC1 -1.494 · PC2 0.3637PC1 0.1517 · PC2 0.2679PC1 -0.5147 · PC2 0.3041PC1 0.2144 · PC2 0.3635PC1 -0.5411 · PC2 0.3019PC1 -0.8059 · PC2 0.3257PC1 -7.004 · PC2 0.1243PC1 -3.174 · PC2 0.2704PC1 0.819 · PC2 0.2815PC1 0.5767 · PC2 0.287PC1 0.5434 · PC2 -0.3275PC1 -1.31 · PC2 -0.4516PC1 1.556 · PC2 -0.1689PC1 0.5178 · PC2 -0.2365PC1 -0.2014 · PC2 -0.2868PC1 0.2775 · PC2 -0.2671PC1 -0.03915 · PC2 -0.3122PC1 -1.039 · PC2 -0.3035PC1 -0.5417 · PC2 -0.3062PC1 -0.1199 · PC2 -0.3658PC1 1.351 · PC2 -0.1908PC1 0.6632 · PC2 -0.1959PC1 0.3054 · PC2 -0.2825PC1 0.4871 · PC2 -0.3208PC1 -0.05688 · PC2 -0.2911PC1 0.5791 · PC2 -0.2449PC1 0.7013 · PC2 -0.2986PC1 -0.4206 · PC2 -0.3074PC1 2.122 · PC2 -0.1438PC1 -1.463 · PC2 -0.4191PC1 -0.8127 · PC2 -0.3632PC1 2.554 · PC2 -0.1208PC1 -0.01286 · PC2 -0.3129PC1 -0.3188 · PC2 -0.335PC1 -0.9182 · PC2 -0.4494PC1 -0.4026 · PC2 -0.4966PC1 -0.5599 · PC2 -0.2223PC1 -0.7184 · PC2 -0.3625PC1 0.6257 · PC2 -0.2473PC1 -0.4529 · PC2 -0.2325PC1 -0.8282 · PC2 -0.3564PC1 0.5896 · PC2 -0.2794PC1 -0.01126 · PC2 -0.3304PC1 -0.6506 · PC2 -0.3651PC1 1.537 · PC2 -0.1603PC1 1.019 · PC2 -0.227PC1 0.1861 · PC2 0.2333PC1 -3.09 · PC2 0.369PC1 1.33 · PC2 0.3278PC1 0.5945 · PC2 0.3636PC1 -0.5726 · PC2 0.354PC1 0.5103 · PC2 0.3918PC1 -0.4343 · PC2 0.3288PC1 -1.623 · PC2 0.4866PC1 -0.7255 · PC2 0.3877PC1 -0.1622 · PC2 0.3198PC1 1.416 · PC2 0.3439PC1 0.3408 · PC2 0.3536PC1 0.9432 · PC2 0.3226PC1 0.189 · PC2 0.2454PC1 -0.2718 · PC2 0.3418PC1 0.1974 · PC2 0.2777PC1 0.5052 · PC2 0.2693PC1 -0.7668 · PC2 0.2745PC1 1.676 · PC2 0.2834PC1 -1.852 · PC2 0.2628PC1 -1.14 · PC2 0.3251PC1 2.309 · PC2 0.3016PC1 -0.3825 · PC2 0.3009PC1 -0.5476 · PC2 0.3075PC1 -1.426 · PC2 0.171PC1 -0.9534 · PC2 0.09936PC1 -0.8798 · PC2 0.3334PC1 -0.7249 · PC2 0.3126PC1 0.244 · PC2 0.263PC1 -0.6213 · PC2 0.3732PC1 -1.099 · PC2 0.2837PC1 0.1784 · PC2 0.2675PC1 -0.391 · PC2 0.2341PC1 -1.013 · PC2 0.2403PC1 1.387 · PC2 0.2957PC1 0.6785 · PC2 0.3054PC1 (88.0%)PC2 (6.1%)216 scores
PCA explained variance0%25%50%75%100%PC1: 88.0% (cumulative 88.0%)1PC2: 6.1% (cumulative 94.1%)2PC3: 4.3% (cumulative 98.3%)3PC4: 0.6% (cumulative 98.9%)4PC5: 0.1% (cumulative 99.0%)5PC6: 0.1% (cumulative 99.1%)6PC7: 0.1% (cumulative 99.1%)7PC8: 0.0% (cumulative 99.2%)8PC9: 0.0% (cumulative 99.2%)9PC10: 0.0% (cumulative 99.2%)10cumulative explained variancePC variancecumulativeprincipal component · cumulative (dashed)

transmittance

X_transmittance · NIR
transmittance spectra-0.20.00.20.40.60.82004006008001,0001,200q05-q95 envelopeq25-q75 envelopemedian spectrummedianq25–q75q05–q95wavelength / nm394.92nm — median 0.009816 (q25–q75 0.005788–0.01526)399.62nm — median 0.01997 (q25–q75 0.01619–0.02454)404.31nm — median 0.01681 (q25–q75 0.0134–0.02138)409.01nm — median 0.01409 (q25–q75 0.01114–0.01805)413.7nm — median 0.01122 (q25–q75 0.008881–0.01456)418.39nm — median 0.009086 (q25–q75 0.006885–0.01225)423.08nm — median 0.00752 (q25–q75 0.005634–0.01027)427.77nm — median 0.006212 (q25–q75 0.004682–0.008562)432.45nm — median 0.005269 (q25–q75 0.003954–0.007373)437.13nm — median 0.004862 (q25–q75 0.003607–0.006854)441.81nm — median 0.005154 (q25–q75 0.003873–0.0073)446.48nm — median 0.006529 (q25–q75 0.004938–0.009135)451.15nm — median 0.008581 (q25–q75 0.006517–0.01155)455.82nm — median 0.009836 (q25–q75 0.00739–0.01302)460.49nm — median 0.01008 (q25–q75 0.007449–0.01311)465.15nm — median 0.009925 (q25–q75 0.007369–0.01295)469.81nm — median 0.009997 (q25–q75 0.007464–0.01311)474.47nm — median 0.01067 (q25–q75 0.00803–0.01389)479.13nm — median 0.01171 (q25–q75 0.008903–0.01524)483.78nm — median 0.01259 (q25–q75 0.00968–0.01642)488.43nm — median 0.0138 (q25–q75 0.01079–0.01811)493.08nm — median 0.01656 (q25–q75 0.01298–0.02141)497.73nm — median 0.02109 (q25–q75 0.0169–0.02712)502.37nm — median 0.02886 (q25–q75 0.02315–0.0356)506.24nm — median 0.03895 (q25–q75 0.03137–0.04661)510.87nm — median 0.05566 (q25–q75 0.04543–0.06547)515.51nm — median 0.0762 (q25–q75 0.06357–0.08883)520.14nm — median 0.1008 (q25–q75 0.08484–0.1164)524.77nm — median 0.1231 (q25–q75 0.1046–0.1414)529.4nm — median 0.1384 (q25–q75 0.1194–0.1608)534.02nm — median 0.1485 (q25–q75 0.1284–0.1733)538.65nm — median 0.1552 (q25–q75 0.134–0.1803)543.27nm — median 0.1598 (q25–q75 0.1383–0.1857)547.88nm — median 0.1643 (q25–q75 0.1424–0.1913)552.5nm — median 0.1675 (q25–q75 0.1459–0.1957)557.11nm — median 0.1673 (q25–q75 0.1462–0.1948)561.72nm — median 0.1635 (q25–q75 0.1424–0.1887)566.32nm — median 0.1556 (q25–q75 0.1354–0.1788)570.93nm — median 0.1443 (q25–q75 0.1255–0.1645)575.53nm — median 0.1326 (q25–q75 0.1151–0.1506)580.13nm — median 0.1231 (q25–q75 0.1077–0.1404)584.72nm — median 0.1157 (q25–q75 0.1015–0.1328)589.32nm — median 0.1109 (q25–q75 0.09636–0.1277)593.9nm — median 0.1079 (q25–q75 0.0935–0.1249)598.49nm — median 0.1059 (q25–q75 0.09215–0.1235)603.08nm — median 0.1033 (q25–q75 0.08997–0.1209)607.66nm — median 0.0984 (q25–q75 0.08501–0.1154)612.24nm — median 0.09116 (q25–q75 0.07849–0.1076)616.82nm — median 0.08481 (q25–q75 0.07258–0.09982)621.39nm — median 0.08016 (q25–q75 0.0684–0.09483)625.96nm — median 0.07836 (q25–q75 0.06636–0.09287)630.53nm — median 0.07778 (q25–q75 0.06573–0.09237)635.1nm — median 0.07523 (q25–q75 0.06365–0.08936)639.66nm — median 0.06745 (q25–q75 0.05727–0.08079)644.23nm — median 0.05631 (q25–q75 0.04771–0.06906)648.78nm — median 0.04682 (q25–q75 0.03895–0.05726)653.34nm — median 0.04153 (q25–q75 0.03411–0.05107)657.89nm — median 0.0353 (q25–q75 0.02844–0.04377)662.44nm — median 0.02625 (q25–q75 0.02019–0.03269)666.99nm — median 0.01789 (q25–q75 0.01314–0.02354)671.54nm — median 0.01327 (q25–q75 0.009873–0.01832)676.08nm — median 0.01158 (q25–q75 0.008444–0.01632)680.62nm — median 0.01255 (q25–q75 0.009239–0.0173)685.16nm — median 0.02203 (q25–q75 0.01623–0.02846)689.69nm — median 0.05239 (q25–q75 0.04289–0.06284)694.23nm — median 0.1078 (q25–q75 0.09141–0.1244)698.75nm — median 0.1722 (q25–q75 0.1511–0.1974)703.28nm — median 0.2314 (q25–q75 0.2099–0.2617)707.81nm — median 0.2817 (q25–q75 0.2599–0.3156)712.33nm — median 0.33 (q25–q75 0.3032–0.3674)716.09nm — median 0.3649 (q25–q75 0.336–0.4065)720.61nm — median 0.4045 (q25–q75 0.3696–0.4477)725.12nm — median 0.4369 (q25–q75 0.3984–0.4784)729.64nm — median 0.463 (q25–q75 0.4215–0.5039)734.14nm — median 0.4827 (q25–q75 0.4375–0.5221)738.65nm — median 0.4956 (q25–q75 0.4484–0.5347)743.15nm — median 0.5048 (q25–q75 0.4566–0.5435)747.65nm — median 0.5138 (q25–q75 0.4623–0.551)752.15nm — median 0.5179 (q25–q75 0.4674–0.5547)756.65nm — median 0.5199 (q25–q75 0.4684–0.5555)761.14nm — median 0.5203 (q25–q75 0.4696–0.5569)765.63nm — median 0.5217 (q25–q75 0.4705–0.5568)770.12nm — median 0.5216 (q25–q75 0.4708–0.558)774.61nm — median 0.5212 (q25–q75 0.472–0.5589)779.09nm — median 0.5198 (q25–q75 0.4724–0.558)783.57nm — median 0.5193 (q25–q75 0.4718–0.5597)788.04nm — median 0.5195 (q25–q75 0.4721–0.5617)792.52nm — median 0.5181 (q25–q75 0.4703–0.5593)796.99nm — median 0.5194 (q25–q75 0.4731–0.5598)801.46nm — median 0.5173 (q25–q75 0.4714–0.5612)805.93nm — median 0.517 (q25–q75 0.4705–0.5604)810.39nm — median 0.5163 (q25–q75 0.4712–0.5596)814.85nm — median 0.5142 (q25–q75 0.4682–0.558)819.31nm — median 0.5116 (q25–q75 0.467–0.5561)823.77nm — median 0.5099 (q25–q75 0.464–0.5544)828.22nm — median 0.5071 (q25–q75 0.4609–0.5482)832.67nm — median 0.5017 (q25–q75 0.4545–0.5436)837.12nm — median 0.4991 (q25–q75 0.4517–0.5412)841.56nm — median 0.4957 (q25–q75 0.4516–0.5378)846.01nm — median 0.4955 (q25–q75 0.4528–0.5387)850.45nm — median 0.498 (q25–q75 0.4517–0.5427)854.88nm — median 0.4894 (q25–q75 0.4464–0.5343)859.32nm — median 0.4827 (q25–q75 0.4401–0.5222)863.75nm — median 0.4674 (q25–q75 0.4248–0.509)868.18nm — median 0.4591 (q25–q75 0.4159–0.4984)872.61nm — median 0.4521 (q25–q75 0.4114–0.4928)877.03nm — median 0.446 (q25–q75 0.4069–0.4899)881.45nm — median 0.4375 (q25–q75 0.4015–0.4791)885.87nm — median 0.4178 (q25–q75 0.383–0.4653)890.29nm — median 0.4072 (q25–q75 0.3729–0.4527)894.7nm — median 0.3943 (q25–q75 0.3604–0.4434)899.11nm — median 0.3848 (q25–q75 0.3484–0.4335)903.52nm — median 0.395 (q25–q75 0.3579–0.4422)907.93nm — median 0.3791 (q25–q75 0.3428–0.4255)912.33nm — median 0.3731 (q25–q75 0.3348–0.4176)916.73nm — median 0.3572 (q25–q75 0.3194–0.404)920.4nm — median 0.3399 (q25–q75 0.3019–0.3849)924.79nm — median 0.331 (q25–q75 0.2948–0.3771)929.18nm — median 0.3104 (q25–q75 0.2701–0.3534)933.58nm — median 0.2901 (q25–q75 0.2555–0.3397)937.96nm — median 0.2838 (q25–q75 0.242–0.3211)942.35nm — median 0.2688 (q25–q75 0.2319–0.3116)946.73nm — median 0.2582 (q25–q75 0.2214–0.3013)951.11nm — median 0.2351 (q25–q75 0.1998–0.2709)955.49nm — median 0.2193 (q25–q75 0.1864–0.2575)959.86nm — median 0.2053 (q25–q75 0.1732–0.2421)964.23nm — median 0.1938 (q25–q75 0.1638–0.228)968.6nm — median 0.1859 (q25–q75 0.1554–0.2145)972.97nm — median 0.1717 (q25–q75 0.1442–0.1973)977.33nm — median 0.1625 (q25–q75 0.1359–0.1888)981.69nm — median 0.151 (q25–q75 0.1275–0.1739)986.05nm — median 0.1428 (q25–q75 0.1204–0.166)990.41nm — median 0.1363 (q25–q75 0.1154–0.1583)994.76nm — median 0.1299 (q25–q75 0.1111–0.1483)999.11nm — median 0.1264 (q25–q75 0.1083–0.1478)1003.5nm — median 0.1248 (q25–q75 0.1069–0.1462)1007.8nm — median 0.121 (q25–q75 0.1061–0.1397)1012.1nm — median 0.1202 (q25–q75 0.105–0.1374)1016.5nm — median 0.1183 (q25–q75 0.1058–0.137)1020.8nm — median 0.118 (q25–q75 0.1041–0.1355)

Sampling

Wavelengths832
Axis range394.9–1021 nm
Mean spacing0.753 nm
Gridirregular
Observations216

Signal & quality

Value range-0.019 – 0.918
Mean range0.00579 – 0.525
Mean level0.2313
Area143.3
PTP0.5193
Noise RMS0.0003376
SNR6.9e+02
SNR dB6e+01 dB
Dynamic range0.519
Smoothness0.009922
Saturated0.0%
X-outliers117

Integrity & artefacts

NaN ratio0.00%
Inf count0
Zero ratio0.00%
Spike count39,354
Spike rate21.95%
Jump count3,942
Jump rate2.20%
Clip fraction0.00%

Shape & reference

Baseline slope0.38895
Curvature RMS0.010064
D1 RMS0.0062052
RMS to mean0.033064
RMS p950.096273
SAM to mean0.051753
SAM p950.080447
Affine offset p950.025291
Affine gain p95 Δ0.3332
Affine residual p950.022089
Xcorr lag p950

Outliers & repeatability

PCA Q p95/median2.7
Hotelling T2 p95/median4.3
Mahalanobis H p95/median2.1
Repeat groups0

Dimensionality (PCA)

Effective rank1.6
PCs → 95% var3
PCs → 99% var4
Top-10 cum. var99.4%
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_reflectance0.231281.00fortValeur 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_curve143.251.00fortValeur 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_peak0.519270.00faibleVariabilité forteSaturationmax(mean_spectrum) - min(mean_spectrum)alert increases when dynamic range is abnormally flat
Amplitude globaleVarianceamplitude.variance0.0394930.00faibleNormal ou hétérogèneMauvais contactvar(X finite)alert increases when variance/dynamic range is abnormally flat
BruitNoise RMSnoise.noise_rms0.00033760.01faibleStableLampe, détecteurmedian MAD(second derivative) * 1.4826 / sqrt(6)alert = noise_rms / signal_scale, saturated at 5%
BruitSNRnoise.snr685.090.00faibleBon signalAcquisitionmean(abs(X)) / noise_rmsalert decreases with SNR dB; >=40 dB is treated as low alert
BruitBandwise SNRnoise.bandwise_snr_min1.28980.94fortZone problématiqueDétecteurmin(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_count39,3541.00fortArtefactsCosmic rays, splicecount robust outliers in second derivativealert follows spike_rate, saturated at 1%
Artefacts locauxSpike rateartefacts.spike_rate22%1.00fortSpectre suspectInterpolationspike_count / (n_samples * (n_features - 2))alert = min(1, spike_rate / 0.01)
Artefacts locauxJump countartefacts.jump_count3,9421.00fortRaccord détecteurSplicecount robust outliers in first derivativealert follows jump_rate, saturated at 1%
Artefacts locauxJump rateartefacts.jump_rate2.2%1.00fortProblème spectralCalibrationjump_count / (n_samples * (n_features - 1))alert = min(1, jump_rate / 0.01)
Artefacts locauxClip fractionartefacts.clip_fraction0.00111%0.00faibleNormalDétecteur saturéfraction of finite cells equal to repeated min/max extremaalert = min(1, clip_fraction / 0.01)
Forme spectraleBaseline slopeshape.baseline_slope0.388951.00fortDériveÉclairementlinear slope of mean_spectrum over normalized axisalert = abs(slope / signal_scale), saturated at 0.5
Forme spectraleCurvature RMSshape.curvature_rms0.0100641.00fortForme inhabituelleFond, splicemedian RMS(second derivative per spectrum)alert = curvature_rms / signal_scale, saturated at 1%
Forme spectraleD1 RMSshape.d1_rms0.00620520.24faiblePlatBiologie ou artefactmedian RMS(first derivative per spectrum)alert = d1_rms / signal_scale, saturated at 5%
Outliers multivariésPCA Q (SPE)outliers.pca_q_ratio2.72090.34faibleConformeArtefact, mélangep95(Q/SPE residual) / median(Q/SPE residual)alert = min(1, pca_q_ratio / 8)
Outliers multivariésHotelling T²outliers.hotelling_t2_ratio4.25690.53moyenExtrême mais cohérentVariabilité naturellep95(Hotelling T2) / median(Hotelling T2)alert = min(1, hotelling_t2_ratio / 8)
Outliers multivariésMahalanobis Houtliers.mahalanobis_h_ratio2.05780.51moyenOutlier globalDomaine différentp95(sqrt(T2)) / median(sqrt(T2))alert = min(1, mahalanobis_h_ratio / 4)
Comparaison à référenceRMS to mean spectrumreference.rms_to_mean_spectrum_p950.0962730.74fortSpectre différentDomain 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_p950.0804470.23faibleSimilaireFond, 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_density4.34221.00fortSous-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_p953.41341.00fortSpectre 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_p950.595371.00fortSpectre atypiqueDiverses causesp95 IsolationForest anomaly score on PCA scoresalert follows structure complexity; raw score is implementation-dependent
X_transmittance PCA score plot-10-505-1.0-0.50.00.51.0PC1 -0.2065 · PC2 -0.453PC1 -3.53 · PC2 -0.7897PC1 -2.003 · PC2 -0.6309PC1 -2.645 · PC2 -0.7204PC1 -0.1165 · PC2 -0.4601PC1 -4.427 · PC2 -0.8518PC1 -0.321 · PC2 0.282PC1 -3.152 · PC2 0.3349PC1 -1.623 · PC2 0.4008PC1 -2.754 · PC2 0.2506PC1 -0.947 · PC2 0.2652PC1 -6.847 · PC2 0.2129PC1 1.908 · PC2 -0.2199PC1 0.1855 · PC2 -0.4279PC1 0.1299 · PC2 -0.4125PC1 0.3829 · PC2 -0.3836PC1 0.2957 · PC2 -0.3924PC1 1.818 · PC2 -0.2224PC1 0.2113 · PC2 -0.4048PC1 1.264 · PC2 -0.2781PC1 -0.04799 · PC2 -0.3678PC1 0.2114 · PC2 -0.3307PC1 1.12 · PC2 -0.2648PC1 1.653 · PC2 -0.2762PC1 -0.07515 · PC2 -0.3629PC1 0.1344 · PC2 -0.3826PC1 0.338 · PC2 -0.3476PC1 0.514 · PC2 -0.3565PC1 -0.4422 · PC2 -0.4237PC1 0.7294 · PC2 -0.3135PC1 1.193 · PC2 -0.28PC1 1.923 · PC2 -0.1824PC1 0.841 · PC2 -0.3497PC1 0.3067 · PC2 -0.3946PC1 -1.298 · PC2 -0.5342PC1 2.187 · PC2 -0.1896PC1 0.6289 · PC2 -0.3175PC1 -0.5371 · PC2 -0.4782PC1 -0.4302 · PC2 -0.4428PC1 -1.281 · PC2 -0.5051PC1 2.622 · PC2 -0.1585PC1 1.533 · PC2 -0.2656PC1 1.602 · PC2 0.381PC1 -0.1752 · PC2 0.319PC1 -0.3226 · PC2 0.3194PC1 -0.08418 · PC2 0.3302PC1 0.008146 · PC2 0.3317PC1 1.598 · PC2 0.4285PC1 -0.1072 · PC2 0.346PC1 0.9736 · PC2 0.4145PC1 -0.5743 · PC2 0.3497PC1 -0.1931 · PC2 0.4124PC1 0.6245 · PC2 0.4638PC1 0.7799 · PC2 0.295PC1 -0.6547 · PC2 0.3953PC1 -0.3074 · PC2 0.3617PC1 -0.1586 · PC2 0.3876PC1 -3.421 · PC2 0.3306PC1 -0.8849 · PC2 0.3616PC1 0.2925 · PC2 0.3405PC1 0.8609 · PC2 0.3509PC1 1.708 · PC2 0.4351PC1 0.1661 · PC2 0.2614PC1 -0.2193 · PC2 0.3207PC1 -2.035 · PC2 0.2934PC1 1.561 · PC2 0.3431PC1 0.2912 · PC2 0.3978PC1 -1.062 · PC2 0.2758PC1 -0.872 · PC2 0.2645PC1 -2.02 · PC2 0.3316PC1 2.291 · PC2 0.3066PC1 1.171 · PC2 0.3653PC1 0.4137 · PC2 -0.3737PC1 0.1196 · PC2 -0.436PC1 -0.7135 · PC2 -0.4971PC1 -0.5562 · PC2 -0.4855PC1 0.9705 · PC2 -0.2699PC1 -0.6014 · PC2 -0.4366PC1 -0.5577 · PC2 -0.4547PC1 -0.06153 · PC2 -0.4258PC1 0.6681 · PC2 -0.2873PC1 -0.7459 · PC2 -0.4558PC1 0.1121 · PC2 -0.3769PC1 0.7516 · PC2 -0.2578PC1 0.3469 · PC2 -0.403PC1 -0.03972 · PC2 -0.4324PC1 0.8914 · PC2 -0.3104PC1 1.45 · PC2 -0.2196PC1 0.08773 · PC2 -0.3794PC1 0.2923 · PC2 -0.3495PC1 1.485 · PC2 -0.2447PC1 0.42 · PC2 -0.3245PC1 0.9281 · PC2 -0.2374PC1 -1.233 · PC2 -0.4501PC1 2.062 · PC2 -0.1518PC1 2.936 · PC2 -0.1092PC1 1.403 · PC2 -0.2935PC1 -1.643 · PC2 -0.5604PC1 -0.2786 · PC2 -0.3859PC1 1.099 · PC2 -0.2853PC1 0.9079 · PC2 -0.2883PC1 0.4412 · PC2 -0.2846PC1 -0.3644 · PC2 -0.4006PC1 -0.01532 · PC2 -0.3543PC1 0.1879 · PC2 -0.3562PC1 -0.01985 · PC2 -0.3951PC1 2.169 · PC2 -0.1724PC1 1.409 · PC2 -0.236PC1 0.1529 · PC2 0.4041PC1 -0.3135 · PC2 0.2866PC1 -1.129 · PC2 0.3207PC1 -0.9878 · PC2 0.2794PC1 0.4943 · PC2 0.3483PC1 -0.9653 · PC2 0.3616PC1 -0.8289 · PC2 0.3557PC1 -0.4141 · PC2 0.3076PC1 -0.003441 · PC2 0.4043PC1 -1.408 · PC2 0.2934PC1 -0.2256 · PC2 0.3628PC1 0.3881 · PC2 0.5329PC1 0.017 · PC2 0.317PC1 -0.6852 · PC2 0.304PC1 0.5028 · PC2 0.3285PC1 1.158 · PC2 0.3625PC1 -0.3961 · PC2 0.3639PC1 -0.07264 · PC2 0.357PC1 1.096 · PC2 0.3869PC1 -0.2699 · PC2 0.3722PC1 0.5612 · PC2 0.4591PC1 -1.886 · PC2 0.3607PC1 1.573 · PC2 0.4366PC1 2.687 · PC2 0.453PC1 1.127 · PC2 0.3891PC1 -2.258 · PC2 0.27PC1 -0.7366 · PC2 0.4228PC1 0.6495 · PC2 0.3676PC1 0.4705 · PC2 0.3492PC1 0.0344 · PC2 0.4046PC1 -0.9468 · PC2 0.3404PC1 -0.4976 · PC2 0.3959PC1 -5.527 · PC2 0.3067PC1 -2.634 · PC2 0.3135PC1 1.657 · PC2 0.3834PC1 0.7856 · PC2 0.3721PC1 0.1556 · PC2 -0.3974PC1 -1.287 · PC2 -0.5142PC1 1.489 · PC2 -0.2055PC1 -1.023 · PC2 -0.4826PC1 -0.6578 · PC2 -0.418PC1 1.059 · PC2 -0.2392PC1 -1.239 · PC2 -0.5135PC1 -1.66 · PC2 -0.4884PC1 -0.8376 · PC2 -0.4341PC1 0.4731 · PC2 -0.3296PC1 0.7285 · PC2 -0.2879PC1 0.3014 · PC2 -0.3321PC1 -0.3481 · PC2 -0.4067PC1 1.097 · PC2 -0.2651PC1 1.174 · PC2 -0.1447PC1 0.4297 · PC2 -0.3255PC1 2.39 · PC2 -0.151PC1 -0.8623 · PC2 -0.3916PC1 1.397 · PC2 -0.2288PC1 -0.5813 · PC2 -0.3656PC1 -0.2224 · PC2 -0.4214PC1 3.59 · PC2 0.02432PC1 1.155 · PC2 -0.2049PC1 1.325 · PC2 -0.1725PC1 0.9369 · PC2 -0.2971PC1 -0.5714 · PC2 -0.4759PC1 0.8777 · PC2 -0.1571PC1 1.212 · PC2 -0.2344PC1 -0.8647 · PC2 -0.4721PC1 -1.82 · PC2 -0.4996PC1 -0.8219 · PC2 -0.4255PC1 0.4824 · PC2 -0.3025PC1 0.8717 · PC2 -0.2629PC1 0.4051 · PC2 -0.2838PC1 1.006 · PC2 -0.2623PC1 0.999 · PC2 -0.2499PC1 -0.3409 · PC2 0.3126PC1 -3.546 · PC2 0.3494PC1 1.064 · PC2 0.3826PC1 -1.193 · PC2 0.4708PC1 -1.436 · PC2 0.2941PC1 0.864 · PC2 0.4027PC1 -1.97 · PC2 0.2843PC1 -2.624 · PC2 0.4129PC1 -1.328 · PC2 0.35PC1 0.2788 · PC2 0.3991PC1 1.768 · PC2 0.4521PC1 -0.262 · PC2 0.3649PC1 -0.4165 · PC2 0.2953PC1 0.7084 · PC2 0.3602PC1 0.7864 · PC2 0.4767PC1 -0.1418 · PC2 0.3145PC1 2.013 · PC2 0.3323PC1 -1.408 · PC2 0.3414PC1 0.716 · PC2 0.3277PC1 -1.207 · PC2 0.3386PC1 -0.6378 · PC2 0.2297PC1 3.2 · PC2 0.4469PC1 0.5827 · PC2 0.3785PC1 0.9175 · PC2 0.4265PC1 0.4428 · PC2 0.3653PC1 -1.288 · PC2 0.2515PC1 0.5397 · PC2 0.4561PC1 1.018 · PC2 0.3659PC1 -1.495 · PC2 0.2013PC1 -2.385 · PC2 0.318PC1 -1.328 · PC2 0.3203PC1 -0.1447 · PC2 0.3389PC1 0.4288 · PC2 0.4332PC1 -0.1217 · PC2 0.3813PC1 0.6304 · PC2 0.3106PC1 0.3844 · PC2 0.3519PC1 (88.8%)PC2 (6.1%)216 scores
PCA explained variance0%25%50%75%100%PC1: 88.8% (cumulative 88.8%)1PC2: 6.1% (cumulative 95.0%)2PC3: 3.9% (cumulative 98.8%)3PC4: 0.4% (cumulative 99.2%)4PC5: 0.1% (cumulative 99.3%)5PC6: 0.0% (cumulative 99.3%)6PC7: 0.0% (cumulative 99.3%)7PC8: 0.0% (cumulative 99.4%)8PC9: 0.0% (cumulative 99.4%)9PC10: 0.0% (cumulative 99.4%)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 3

treatment

target · categorical
treatment classescc: 8080nn: 6868ss: 6868
n / missing216 / 0
Classes3
Balance (entropy)1
Imbalance ratio1
Top classc (80)

par_condition

target · categorical
par_condition classesPAR1000PAR1000: 108108PAR300PAR300: 108108
n / missing216 / 0
Classes2
Balance (entropy)1
Imbalance ratio1
Top classPAR1000 (108)

day_after_stress

target · categorical
day_after_stress classes44: 727277: 727222: 606000: 1212
n / missing216 / 0
Classes4
Balance (entropy)0.9
Imbalance ratio6
Top class4 (72)

Metadata 2

replicate

metadata · categorical
replicate classesr5r5: 4242r4r4: 3838r1r1: 3636r2r2: 3636r3r3: 3636r6r6: 2626r7r7: 22
n / missing216 / 0
Classes7
Balance (entropy)0.93
Imbalance ratio21
Top classr5 (42)

leaf_sample

metadata · categorical
leaf_sample classesL2L2: 104104L1L1: 5454L1xL1x: 52521x1x: 2222: 22L1bisL1bis: 22
n / missing216 / 0
Classes6
Balance (entropy)0.65
Imbalance ratio52
Top classL2 (104)
Constant metadata 14
  • doi_dataset10.5281/zenodo.12800065
  • licenseCreative Commons Attribution 4.0 International
  • license_idcc-by-4.0
  • license_urlhttps://creativecommons.org/licenses/by/4.0/
  • rights_statusexplicit_open
  • usage_scopepublic_reuse_possible
  • public_release_allowedTrue
  • spectroscopy_typeVis-NIR
  • axis_unitnm
  • axis_min394.9
  • axis_max1021
  • n_points_original832
  • decodermat-io/matio
  • notesMATLAB string labels decoded with matio load_from_mat(raw_data=False).

Alignment

Alignment levelsample
Sample id availableyes
Samples216
Observations (total)648
Reps per samplemin 1 · mean 1 · max 1

Provenance & citation

ContributorTomato leaf stress spectra
Origin · zenodo [open]https://zenodo.org/records/12800065 — Zenodo
Origin · url [open]https://creativecommons.org/licenses/by/4.0/
Origin · script [manual]source_to_standard.py — standardization script (maintainer-only)

Governance & integrity

Tierpublic
LicenseCC-BY-4.0
Permitted useResearch and benchmarking.
Access policyOpen per source license.
RedistributionZenodo record reports CC BY 4.0.
Content version1.0.0
Schema / protocol2.0
Content hashd9b3f6eb2be966f8…
Processing hash08496f5f7b8fea6a…
Metadata hash034b75fb3d9be352…

Load this dataset

# pip install nirs4all-datasets
from nirs4all_datasets import get

ds = get("zenodo_tomato_leaf_stress_visnir_matio")            # DOI-pinned, checksum-verified, cached
X, y = ds.x(), ds.y()
print(X.shape, y.shape)
card.jsoncroissant.jsonIdentity metadata only — the dataset bytes live at the origin / DOI.