Quasi-likelihood estimating equations (quasi-score function) for the estimation is as follows $$\sum_i\frac{\partial{\mu_i^{'}}}{\partial{\beta}}V_i^{-1}(y_i-\mu_i)=0.$$
The $\frac{\partial{\mu_i^{'}}}{\partial{\beta}}=X_{n,p+1}$ matrix. The extra column is to accommodate for intercept $\alpha$. However $V_{p,p}$ is a $p\times p$ matrix. This creates incompatibility issues,
$X$ has $p+1$ columns:
[,1] [,2] [,3] [,4]
1 1 26.5 104.9
1 1 26.5 123.0
1 1 26.5 140.9
1 1 26.5 160.0
1 2 28.3 3.3
1 2 28.3 13.2
.
.
.
.
$V$ has $p$ columns:
[,1] [,2] [,3]
11.17 0.56 1.18
0.56 11.34 5.17
1.18 5.17 6.24
How do I solve this problem, what is the correct way to incorporate intercept $\alpha$ . Any help is greatly appreciated, Thanks.
