One weight function, two distances | Euclidean distance | Covariance-based distance | The three neighborhoods side by side | When the covariance-based index helps | When the covariance-based index fails | Choosing in practice | References

The algorithm | Regression: fat-free water content of meat from NIR spectra | Comparison with global PLS | Similarity indexes | Classification | Using the engine directly | Robust and sparse local models | Engine arguments | References

Can I trust this prediction? | Theoretical limits and percentile ranks | Which bands drive the predictions? | Which reference values are suspect? | Which wavelengths are selected? | Customizing the plots | References

Robust LW-PLS | Why | Example: wrong reference values | Sparse LW-PLS | Sparse and robust | References

Baseline correction | Normalization | Other steps | A model for potassium

Detector saturation | Electron density | Temperature | Atomic data | Line intensities | Boltzmann and Saha-Boltzmann plots | Is the plasma in LTE? | Self-absorption | Calibration-free composition | Summary | References

Identifying the lines | Correcting the wavelength axis | Line profiles | Fitting a single line | Overlapping lines | Fitting every spectrum | Integrating instead of fitting | Summary

Line intensities | A calibration curve for potassium | Trace elements: weights and limits of detection | A multivariate model and its figures of merit | Summary

Data | Principal component analysis | Hotelling’s T-Squared Statistic | Calculate using the first k components | Calculate using the cumulative variance threshold | Hotelling’s T-Squared Ellipse | Calculating the semi-axes | Calculating the x-y coordinates | Hotelling’s T-Squared Ellipsoid | Calculating the x-y-z coordinates

Coordinate points | Group-based confidence ellipses

Coordinate points | Grouping