DOE and Multivariate Modeling Techniques for Material Selection ( 2019-EU-45MP-092 )
Feb 11, 2019 4:34 PM
| Last Modified: Mar 14, 2019 2:22 AM
Level: Intermediate Job Function: Analyst / Scientist / Engineer Bill Worley, JMP Senior Systems Engineer, SAS Mahmoud Hammoud, JMP Senior Systems Engineer, SAS
This is an update on a talk given by Silvio Miccio at Discovery Summit Europe 2017. This is truly a revolutionary technique that scientists and engineers should be embracing in their everyday work to improve products and processes. The presenters will demonstrate the use of principal components and covariate DOE for the selection of test options based on their physical, chemical and/or biological properties. Partial least squares along with generalized regression techniques will then be used to build predictive models based on the properties of note. These truly groundbreaking models will then be viable as long as the properties of the candidate materials do not change. After the model is built, any new candidate materials will only need to have their physical properties added to the candidate table to see where they fall in the predictive model space via multidimensional scaling and other multivariate techniques.