Level: Intermediate
In recent years, PLS (Partial Least Squares) regression has been used to develop predictive models using spectral data. This presentation will demonstrate how a new approach using the Functional Data Explorer (FDE) and the covariate features of Design of Experiments (DOE) can be used to develop better predictive models with less spectral data. This new method uses one-quarter to one-third of the data previously used to develop predictive models. We will also introduce the new Model-Driven Multivariate Control Chart (MDMCC) platform in JMP 15 as a way to analyze spectral data.
Prior to joining the JMP team, Bill Worley spent 14 years at Procter & Gamble, the last five of which were as the technology leader for the Modeling and Simulation group, where he educated and supported the use of JMP. Worley is a chemist and holds a Master's degree in Chemistry from the University of Cincinnati.
