Custom Designs for Experiments

This course introduces a state-of-the-art approach to designing industrial laboratory experiments that is based on the latest statistical theory and numerical methods. Advances in computer algorithms and hardware make this approach, once considered exotic and the domain of a few experts, available to everyone for all experiments. In this course you will learn how to use new ways to design experiments that fit the purpose and circumstance of any test and how to use custom designs to extend the underlying design principles to any problem. You will apply the principles behind such choices as model terms and sample size..

Voraussetzungen
No experience with design of experiments is required, but students should be comfortable using JMP and should understand ANOVA and regression.

Zielgruppe
Anyone involved in laboratory experimentation and who wants to learn about powerful new ways to design experiments that exactly fit the purpose and circumstances of any test

Module
JMP Software

Kursinhalte

• Introducing Custom Design
  • familiar results (reproducing classic designs)
  • motivation for unfamiliar solution (everyday situations are not handled by classic designs)
  • optimal design background
  • advantages and benefits
• Accommodating the Real World
  • combinations of different kinds of factors (any kind of factor may be included)
  • constrained combinations of factor levels (disallowed regions or combinations)
  • reusing existing runs (augment existing data with new, complimentary runs)
  • constrained randomization (when some factors are hard to change every run)
  • analyzing multiple responses (ways to find best settings to satisfy more than one goal)
• Not Uncommon Complications (Optional)
  • covariate factors (design with immunity to effect of covariate)
  • design for nonlinear models

Referent
This course will be held by Mark Bailey, Ph.D., Analytical Training Consultant, Education Division, SAS.

zurück...