Skip to content

Keyword index & glossary

You are here: Reference & index → Keyword index & glossary

Questions this page answers - Where do I find a concept (preprocessing method, test, model, metric, workflow, CLI command)? - Which docs and API pages explain it?

Spectral preprocessing

  • ALSBaseline (ALS baseline correction) — removes fluorescence/sloping background. See: ftir_raman_preprocessing.md, api/preprocessing.md#foodspec.preprocess.baseline.ALSBaseline.
  • RubberbandBaseline — convex-hull baseline for concave backgrounds. See: ftir_raman_preprocessing.md, api/preprocessing.md#foodspec.preprocess.baseline.RubberbandBaseline.
  • PolynomialBaseline — low-degree baseline fit. See: ftir_raman_preprocessing.md, api/preprocessing.md#foodspec.preprocess.baseline.PolynomialBaseline.
  • SavitzkyGolaySmoother (SavGol) — noise reduction preserving peaks. See: ftir_raman_preprocessing.md, api/preprocessing.md#foodspec.preprocess.smoothing.SavitzkyGolaySmoother.
  • MovingAverageSmoother — simple denoising; may broaden peaks. See: ftir_raman_preprocessing.md, api/preprocessing.md#foodspec.preprocess.smoothing.MovingAverageSmoother.
  • Vector/Area/Max normalization — scales spectra to unit norm/area. See: ftir_raman_preprocessing.md, api/preprocessing.md#foodspec.preprocess.normalization.VectorNormalizer.
  • SNVNormalizer (Standard Normal Variate) — mean/std per spectrum to reduce scatter. See: ftir_raman_preprocessing.md, api/preprocessing.md#foodspec.preprocess.normalization.SNVNormalizer.
  • MSCNormalizer (Multiplicative Scatter Correction) — corrects additive/multiplicative scatter via reference. See: ftir_raman_preprocessing.md, api/preprocessing.md#foodspec.preprocess.normalization.MSCNormalizer.
  • InternalPeakNormalizer — normalize to a stable internal band/window. See: ftir_raman_preprocessing.md, api/preprocessing.md#foodspec.preprocess.normalization.InternalPeakNormalizer.
  • DerivativeTransformer (derivatives) — Savitzky–Golay derivatives (1st/2nd). See: ftir_raman_preprocessing.md, api/preprocessing.md#foodspec.preprocess.derivatives.DerivativeTransformer.
  • AtmosphericCorrector (FTIR) — remove water/CO₂ contributions. See: ftir_raman_preprocessing.md, api/preprocessing.md#foodspec.preprocess.ftir.AtmosphericCorrector.
  • SimpleATRCorrector (FTIR) — heuristic ATR depth correction. See: ftir_raman_preprocessing.md, api/preprocessing.md#foodspec.preprocess.ftir.SimpleATRCorrector.
  • CosmicRayRemover (Raman) — remove spike artifacts. See: ftir_raman_preprocessing.md, api/preprocessing.md#foodspec.preprocess.raman.CosmicRayRemover.
  • RangeCropper — crop to target wavenumber window. See: ftir_raman_preprocessing.md, api/preprocessing.md#foodspec.preprocess.cropping.RangeCropper.

Features and ratios

  • PeakFeatureExtractor / detect_peaks — peak heights/areas near expected bands. See: workflows/oil_authentication.md, api/features.md#foodspec.features.peaks.PeakFeatureExtractor.
  • integrate_bands — integrate intensity over defined bands. See: workflows/mixture_analysis.md, api/features.md#foodspec.features.bands.integrate_bands.
  • RatioFeatureGenerator / compute_ratios — compute band/peak ratios (e.g., 1655/1742). See: workflows/oil_authentication.md, api/features.md#foodspec.features.ratios.RatioFeatureGenerator.
  • Fingerprint similarity (cosine/correlation) — spectral similarity matrices. See: hyperspectral_tutorial.md, api/features.md#foodspec.features.fingerprint.

Statistical tests

  • t-tests (independent/paired/one-sample) — compare means. See: stats_tests.md.
  • ANOVA / MANOVA — multi-group mean differences. See: stats_tests.md.
  • Mann–Whitney U / Kruskal–Wallis / Wilcoxon / Friedman — non-parametric comparisons. See: stats_tests.md.
  • Correlation (Pearson/Spearman) / simple regression — associations and trends. See: stats_tests.md.

Machine learning models

  • Logistic regression / Linear SVM / PLS-DA — linear classifiers. See: ml/models_and_best_practices.md, api/chemometrics.md#foodspec.chemometrics.models.
  • RBF SVM / k-NN / Random Forest (RF) / Gradient Boosting — nonlinear classifiers. See: ml/models_and_best_practices.md, api/chemometrics.md#foodspec.chemometrics.models.
  • PCA / clustering — unsupervised exploration/visualization. See: chemometrics_guide.md, api/chemometrics.md#foodspec.chemometrics.pca.
  • Conv1DSpectrumClassifier (1D CNN) — optional deep model. See: advanced_deep_learning.md, api/chemometrics.md#foodspec.chemometrics.deep.Conv1DSpectrumClassifier.
  • Mixture models (NNLS, MCR-ALS) — estimate component fractions. See: workflows/mixture_analysis.md, api/chemometrics.md#foodspec.chemometrics.mixture.

Metrics and validation

  • Accuracy, Precision, Recall, F1 (macro/micro), ROC-AUC, Confusion matrix — classification metrics. See: metrics/metrics_and_evaluation/, api/metrics.md.
  • R², RMSE, MAE, Residuals — regression/mixture metrics. See: metrics/metrics_and_evaluation/, api/metrics.md.
  • Cross-validation (CV) — k-fold, stratified CV for models. See: metrics/metrics_and_evaluation/, api/chemometrics.md#foodspec.chemometrics.validation.

Workflows

  • Oil authentication — classify oils/adulteration. See: workflows/oil_authentication.md, api/workflows.md#foodspec.apps.oils.
  • Heating degradation — ratios vs time/temperature. See: workflows/heating_quality_monitoring.md, api/workflows.md#foodspec.apps.heating.
  • Mixture analysis (NNLS/MCR-ALS) — estimate fractions. See: workflows/mixture_analysis.md, api/chemometrics.md#foodspec.chemometrics.mixture.
  • QC / Novelty detection — one-class scoring. See: qc_tutorial.md, api/workflows.md#foodspec.apps.qc.
  • Hyperspectral analysis — ratio/cluster maps. See: hyperspectral_tutorial.md, api/datasets.md#foodspec.core.hyperspectral.HyperSpectralCube.
  • Protocol benchmarks — standardized evaluation. See: protocol_benchmarks.md, api/workflows.md#foodspec.apps.protocol_validation.
  • Domain templates (meat/microbial) — adapt oil workflow to other domains. See: domains_overview.md, meat_tutorial.md, microbial_tutorial.md, api/workflows.md#foodspec.apps.meat, api/workflows.md#foodspec.apps.microbial.

CLI commands

  • about — version/info. See: cli.md.
  • csv-to-library / preprocess — build libraries, preprocess raw data. See: cli.md.
  • oil-auth / heating / qc / domains — workflow commands. See: cli.md.
  • mixture / hyperspectral — mixture and hyperspectral utilities. See: cli.md.
  • protocol-benchmarks — protocol runs. See: cli.md.
  • model-info — inspect saved model metadata. See: cli.md.

See also - Metrics & evaluation - workflows/oil_authentication.md