slc {riv} | R Documentation |
Finds the robust S-estimator of multivariate location and covariance with a high breakdown point based on Tukey's biweight function.
slc(x, nsamp = 500, bdp = 0.5)
x |
a matrix or data frame of data values, say of dimension n x p. The rows represent observations and the columns represent variables. |
nsamp |
the number of random p-subsets considered to compute the S-estimator (default = 500). |
bdp |
breakdown point value of the S-estimator, must be 0.15,0.25 or 0.5 (default). |
A list with components
location |
vector of the estimated multivariate location. |
covariance |
matrix of the estimated covariance. |
distances |
vector of robust Mahalanobis distances versus location and covariance. |
scale |
distance of the scale estimates. |
c |
constant of the Tukey's biweight function. |
Kaufmann B. beat.kaufmann@epfl.ch
Cohen-Freue G.V. gcohen@stat.ubc.ca
Zamar R.H. ruben@stat.ubc.ca
LOPUHAÄ,H.P. (1989) On the Relation between S-estimators and M-estimators of Multivariate Location and Covariance. Ann. Statist. 17 1662-1683.
RUPPERT,D. (1992). Computing S Estimators for Regression and Multivariate Location/Dispersion. J. Comput. Graph. Statist. 1 253-270.
library(MASS) ## load the mortality data-set (62 Alaskan earthquake observations) data(mortality) slc.res <- slc(mortality,nsamp=100,bdp=0.25) ## plot of the Mahalanobis Distances plot(slc.res[[3]],xlab="Observation",ylab="Mahalanobis Distance",main="The Mahalanobis Distance of each observation") ## simulation of a multivariate data-set library(MASS) x <- mvrnorm(100,c(0,0,0),matrix(c(1,0,0,0,1,0.5,0,0.5,1),ncol=3)) slc(x,nsamp=50,bdp=0.15)