The Statistical Processing of Analytical Chemistry Data
Chemometrics Applications (PLS, PCR, PCA or MLR) has been applied to NMR data to obtain predictive models that can predict parameters of interest for process control in the refining, oil exploration, chemical, and food industries.
The chemometrics program used to generate the models for the online applications is Galactic Grams Version 5.0 with the PLS/IQ package. This Galactic software was used for PLS and PLR analyses.
The Pirouette Chemometrics package by Infometrix Inc. is also available for PLS and other applications. Matlab software by Eigenvector research can also be used to perform more cutting edge applications such as non-linear PLS, genetic algorithms, neural nets, ridge regressions, etc. The MLR analyses were performed using the “Analyse-It” Statistical package add-on for Excel.
On the following pages one can find a glossary of terms that are used when describing the comparison of the NMR predictions to the plant laboratory results being used in the chemometrics application validation process: