Résumé :
|
The flexibility of projection methods have made them useful also for the analysis and modelling of messy and complicated data; these methods are increasingly used in a wide range of industrial applications in PARD (processes, administration, research, and development). These include process monitoring and early fault detection and classification, process analytical technology (PAT), data mining and integration, quality control, and relationships between chemical composition/structure and chemical/biological properties. Multivariate calibration, which models the relationship between a multitude of signals measured on various samples and the concentrations of constituents of these samples, is another show-case of MVDA and projections.
|