I am using optFederov, from package AlgDesign, to create D-Optimial designs. However, I am a bit confused on what measure to use as efficiency measure. I'd like to use D-Optimal but I don't know which one is exactly. In the vignette of the package they explain that the determinant of the design is 3.675919 and the optimum approximate theory D is 3.8 making this design about 97% efficient.
1.- But how does he come up with this 3.8? How can I compute it?
2.- When and why should I use D, Ge and Dea?
accoring to the manual D is:
The kth root of the generalized variance: det(M)^(1/k), where det(M) is the determinant of the normalized dispersion matrix M – i.e. M=Z'Z/n, where Z=X[rows,]