How to reconcile prediction profiler confidence intervals with parameter estimates for mixture design
I have an mixture experiment with four components. It's a custom design of 16 runs with no replication that started with a model that included interactions of two factors and three components. ...