In: Statistics and Probability
Consider the general linear model ? = ?? + ?. Use matrix algebra to show that ?̂ is an unbiased estimator of ?.
the last ? has bar
The general linear model is:
is the least square estimate of
The least squre estimate is the one which minimizes the residual errors
Let
(X' is transpose of X)
((AB)' = B'A')
Since, we want the least square esimate of , it should satisfy the following:
Consider equation (2)
(Taking tanspose and cancelling the factor 2)
(Premultiplying both sides by )
Hence,
We have,
for all
Now, would be an unbiased estimator of if
We have,
(E(Y) = X)
Hence, is an unbiased estimator of . (Proved)