The design matrix X of the balanced linear model y=Xβ + ε can be expressed in terms of Kronecker product of matrices of 0's and 1's. Also if we use the properties of the Moore-Penrose inverse, the projection matrix P??=X(X'X)X' can be computed wit...
The design matrix X of the balanced linear model y=Xβ + ε can be expressed in terms of Kronecker product of matrices of 0's and 1's. Also if we use the properties of the Moore-Penrose inverse, the projection matrix P??=X(X'X)X' can be computed without computing the generalized inverse (X'X) of X'X. This can reduce a lot of the computational works and computational storages. In this paper, we derive an iterative method for finding the projection matrix of the balanced linear model with interaction effects or not.