powered by
From a QR object, compute the inverse matrix which is implicitely (but not explicitly!) used to solve the underlying least squares problem.
qr.rtr.inv(qr)
\"qr\" object, typically resulting from qr(.).
qr(.)
The \(p \times p\) matrix \((R'R)^{-1}\) or equivalently, the inverse of \(X'X\) (i.e. t(X) %*% X in R).
t(X) %*% X
qr, qr.R, backsolve.
qr
qr.R
backsolve
# NOT RUN { (h3 <- 1/outer(0:5, 1:3, "+")) rtr <- qr.rtr.inv(qr(h3)) all.equal(c(rtr %*% 1:3), solve(crossprod(h3), 1:3)) # }
Run the code above in your browser using DataLab