Internal function that estimates the central subspace. This will not typically be called directly by the user. dr.fit.y is the generic name of a function that returns the response for the particular dimension reduction regression object. Typically, this just returns a centered version of the response, but for phdres it returns OLS residuals.
dr.fit(object, numdir=4, tol=1e-07, ...) dr.fit.y(object)
object |
dimension reduction regression object |
numdir |
maximum number of dimensions to consider |
tol |
tolerance passed to singular value decomposition |
... |
other arguments passed to dr.fit.M |
evectors |
ordered eigenvectors that describe the estimates of the dimension reduction subspace |
evalues |
ordered eigenvalues |
numdir |
number of eigenvalues |
raw.evectors |
eigenvectors of the rotated data |
decomp |
which decomposition was used? |
Sanford Weisberg <sandy@stat.umn.edu>