BUCHI NIRCal - Manual (Page 191)

Chemometrics
NIRCal 5.5 Manual, Version A 191
3.18.62 Property Residuum
Description The difference between the original property and the predicted property.
Use See where (property, spectra) the major differences are.
Method PCR / PLS / Cluster (CLU) / SIMCA
Matrices ID 11
Tip
Details
Related Topic Original Property , Predicted Property
Qualitative calibrations
The predicted property value depends on the cluster, in which the spectrum is located. For the
property residuum, only a value of 0 is acceptable. Property residuum zero means all spectra are in
the correct cluster.
Property residuum +1 means that a spectrum is outside any cluster: it is not identified.
Property residuum
–1 means that a spectrum is in a wrong cluster: it is false identified.
Quantitative calibrations
The values should be as small as possible. Property residuum zero means, the predicted NIR value is
exactly the same, as the referring lab result. An acceptable range is equal to the standard deviation of
the lab method.
3.18.63 Property Score Diffusion
Description Property Score Diffusion is the scaled disorder, number of different property
spectra according the scores of different PC per property. The scaling is
made by the number of spectra of a property.
Use Useful for secondary PCs selection in higher dimensional space (> 3).
Method Cluster (CLU)
Matrices ID 58
Tip It is not depending on the size of the score scattering, only the number of
interchanging of the properties are important. Shows, which property is
mixed and shuffeled with other property spectra. This plot is very informative
in Top View. Dark means low diffusion, white means high diffusion. Select
the PCs, by which more than one dark area (property) appears. Two dark
areas means: these two properties are separated with that PC. Positioning
the mouse over the area, the property name is visible in the Status bar.
Details Small disorder values indicate a clear separation for the property of a given
PC.
Property Score Diffusion is calculated only for the C-Set spectra.
Related Topic Scores , Secondary PCs , Property Score Disorder
Algorithm (unpublished) by Roman Bossart 18. June 1998