dists.2frames {calibrator} | R Documentation |
Distance between points specified by rows of two matrices, according to a positive definite matrix. If not specified, the second matrix used is the first.
dists.2frames(a, b=NULL, A=NULL, A.lower=NULL, test.for.symmetry=TRUE)
a |
First dataframe whose rows are the points |
b |
Second dataframe whose rows are the points; if NULL ,
use a |
A |
Positive definite matrix; if NULL , a value for
A.lower is needed. If a value for A is supplied, use
a clear but possibly slower method |
A.lower |
The lower triangular Cholesky decomposition of
A (only needed if A is NULL ).
If a value for A.lower is specified, this means that a
relatively opaque but possibly faster method will be used. The time
saving ought to be negligible unless nrow(a) (or
nrow(b) if supplied), is huge. Note that this option does
not test for symmetry of matrix A |
test.for.symmetry |
Boolean, with default TRUE meaning
to calculate all element arrays (elegantly), and FALSE
meaning to calculate only the upper triangular elements (using
loops), which ought to be faster. The value of this argument should
not affect the returned value, up to numerical accuracy |
Robin K. S. Hankin
data(toys) dists.2frames(a=D2.toy,A=diag(2)) A <- diag(2) + matrix(0.2,2,2) A.lower <- t(chol(A)) jj.1 <- dists.2frames(a=D2.toy, A=A, test=TRUE) jj.2 <- dists.2frames(a=D2.toy, A=A, test=FALSE) jj.3 <- dists.2frames(a=D2.toy, A.lower=A.lower, test=FALSE) jj.4 <- dists.2frames(a=D2.toy, A.lower=A.lower, test=TRUE)