Week 10 Flashcards
1
Q
Alternative form of blocking model
A
2
Q
Least squares normal equations for alt blocking model in matrix form
A
3
Q
Orthogonal blocking
A
4
Q
General approach to arrange 2m design in k = 2q blocks of size 2m-q
A
1) choose q independent factorial effects (columns) for the defining blocks (typically higher order interactions)
As below,
Then at end:
B1B2 = 1256
B1B3 = 1234
B2B3 = 3456
B1B2B3 = 14678
5
Q
I optimality
A
Minimises average prediction variance
Used more so when predictions are goal
6
Q
Variance invariance matrix for response vector Y given by
A
7
Q
Generalise least squares and variance of
A
GLS