Posts
-
Linear regression part four - Finding an unbiased estimate of the variance in the linear model.
(Here, I borrow heavily from Christensen, Plane Answers to Complex Questions.) Consider again the linear model
-
Linear regression part three - if the mean is linear, the quadratic loss minimizing line is an unbiased estimate of that mean. A normal response yields normality in the quadratic loss minimizer.
In the last two posts, I first found the line that minimizes quadratic loss over a fixed data set. That line, again, is given by the vector:
-
Linear regression part two - finding the covariance of the quadratic loss minimizing line under homoskedasticity.
Last time, I began with a graph.
-
Linear regression part one - linear quadratic loss minimizer.
If there is one thing statisticians are good at, it is drawing straight lines. (Peter Mueller deserves the credit for that – he said it in a lecture once.)
subscribe via RSS