mahalanofix {fpc} | R Documentation |
Computes the vector of (classical or robust)
Mahalanobis distances of all points of x
to the center of the points indexed (or weighted)
by gv
. The latter also determine
the covariance matrix.
Thought for use within fixmahal
.
mahalanofix(x, n = nrow(as.matrix(x)), p = ncol(as.matrix(x)), gv = rep(1, times = n), cmax = 1e+10, method = "ml") mahalanofuz(x, n = nrow(as.matrix(x)), p = ncol(as.matrix(x)), gv = rep(1, times=n), cmax = 1e+10)
x |
a numerical data matrix, rows are points, columns are variables. |
n |
positive integer. Number of points. |
p |
positive integer. Number of variables. |
gv |
for mahalanofix
a logical or 0-1 vector of length n . For mahalanofuz a
numerical vector with values between 0 and 1. |
cmax |
positive number. used in solvecov if
covariance matrix is singular. |
method |
"ml" , "classical" ,
"mcd" or "mve" . Method to compute the covariance
matrix estimator. See cov.rob , fixmahal . |
solvecov
is used to invert the covariance matrix. The methods
"mcd"
and "mve"
in mahalanofix
do not work properly
with point constellations with singular covariance matrices!
A list of the following components:
md |
vector of Mahalanobis distances. |
mg |
mean of the points indexed by gv , weighted mean in
mahalanofuz . |
covg |
covariance matrix of the points indexed by gv ,
weighted covariance matrix in mahalanofuz . |
covinv |
covg inverted by solvecov . |
coll |
logical. If TRUE , covg has been
(numerically) singular. |
Methods "mcd"
and "mve"
require library lqs
.
Christian Hennig chrish@stats.ucl.ac.uk http://www.homepages.ucl.ac.uk/~ucakche/
x <- c(1,2,3,4,5,6,7,8,9,10) y <- c(1,2,3,8,7,6,5,8,9,10) mahalanofix(cbind(x,y),gv=c(0,0,0,1,1,1,1,1,0,0)) mahalanofix(cbind(x,y),gv=c(0,0,0,1,1,1,1,0,0,0)) mahalanofix(cbind(x,y),gv=c(0,0,0,1,1,1,1,1,0,0),method="mcd") mahalanofuz(cbind(x,y),gv=c(0,0,0.5,0.5,1,1,1,0.5,0.5,0))