a formula object, specifying the group variable and feature
variables, with the group variable on the left of a `~'
operator, and the feature variables, separated by +
operators, on the right. If
data
is given, all names used in the formula should be defined as variables
in the data frame.
OPTIONAL ARGUMENTS:
data
a data frame in which to interpret the variables named in
formula. Default is the calling
frame. This is commonly refered to as the training data for the
discriminant function.
family
a
family.discrim object. Currently,
only the
Classical constructor with
its argument
cov.structure equal to
"homoscedastic" or "heteroscedastic" is implemented.
frequencies
a vector of observation frequencies.
na.action
a function to filter missing data. This is applied to the model
frame after any subset argument has been used. The default
na.exclude is to delete the
observation if any missing values are found. A possible alternative
is
na.fail, which generates an
error if any missing values are found.
subset
expression specifying a row subset of the data to be used in the
fit (training data). This can be a logical vector, or a numeric
vector indicating which observation numbers are to be included, or
a character vector with the row names to be included. All
observations are included by default.
prior
a character string or numerical vector specifying the prior
knowledge of the mixing proportions of each group. The acceptable
strings are as follows: "proportional", group of proportions
are the number of observations from each group divided by the
total number of observations; "uniform", group proportions are one over
the number of groups; "none", exclude the mixing proportion from
the discriminant function. If
prior
is a numerical vector, it must have a length equal to the number
of groups and its elements must be positive and sum to one.
method
numerical method to be used to decompose the feature matrix of
covariances. The choices are "svd", singular value decomposition
(the default) and "choleski", Choleski decomposition.
singular.tol
tolerance for determining existing linear dependencies among
the feature vectors.
estim
the robust estimator used by covRob. The choices are: "mcd" for the Fast
MCD algorithm of Rousseeuw and Van Driessen, "donostah" for the
Donoho-Stahel projection based estimator, "M" for the constrained M
estimator provided by Rocke, "pairwiseQC" for the quadrant correlation
based pairwise estimator, and "pairwiseGK" for the Gnanadesikan-Kettenring
pairwise estimator. The default "auto" selects from "donostah", "mcd", and
"pairwiseQC" with the goal of producing a good estimate in a resonable
amount of time.
cov.control
a list of control parameters to be used in the numerical algorithms. See
covRob.control for the possible control parameters and their default
settings.
VALUE:
an object of class
discRob that contains the
robust discriminant function
SEE ALSO:
,
.
EXAMPLES:
discRob(Group ~ ., data = hemo.cont, family = Classical("heter"))