QR decomposition, constraining the R matrix to have non-negative diagonal entries.
qr2(X)A matrix of dimension \(n\) by \(p\) where \(n \ge p\)
Q An \(n\) by \(p\) matrix with orthonormal columns.
R A \(p\) by \(p\) upper-triangular matrix with non-negative
diagonal elements.
This function is almost a wrapper for qr(), qr.R(), and
qr.Q(), except it constrains the diagonal elements of R to be
non-negative. If X is full rank with fewer columns than rows, then
this is sufficient to gaurantee uniqueness of the QR decomposition
(Proposition 5.2 of
Eaton (1983)).
qr, qr.Q, and
qr.R for the base methods on the obtaining the QR
decomposition. lq for the related LQ decomposition.