 Posted Wednesday, May 14, 2008 11:33 PM
|
|
|
|
| Suppose I wish to perform a multivariate linear regression, say y = Xb + (noise). I can use the QR.Solve to tell me the correct "b" coefficients. But I need the inverse of X'X to get the covariance matrix of the b's. How can I EFFICIENTLY compute Inverse(X'X) from the QR factorization of X?
|
|
|
|