discrim object.
discrim object
and statistical tests of the p-variate training data used to compute
the discriminant function.
summary.discrim(object, ...)
discrim object constructed by the
discrim function.
summary method.
summary.discrim object.
Objects of this class have the following methods:
The structure of a
summary.discrim object is a list with the following data members.
discrim object,
object.
method
options of the
predict method, where the default is "plug-in".
See
predict.discrim for details.
The error rate estimates are discussed under the
CV data member.
classify data member.
object is fitted without weights using the spherical, homoscedastic, group
spherical, or heteroscedastic covariance models. See
crossvalidate.discrim for
details on cross-validation.
Nonparametric estimates of the conditional error rates are estimated
based on miss classified counts and base on posterior probabilities.
If the prior probabilities are not proportional to the sample sizes, then stratified
posterior error rates are estimated also. As seen in the example below, it is possible
for the posterior error rate estimates to be negative.
anova.discrim for the
training data used to compute the discriminant function.
These statistics are only computed if the spherical or homoscedastic covariance
models are used. See
anova.discrim for details.
multicomp.discrim. These include
Hotelling's T squared statistics for all pairs of groups in the training
data. For each significant T squared statistic
confidence intervals are computed for the differences between
means of the p-variate feature vectors of the two groups.
anova.discrim for details).
ks.gof for details.
Computes various statistics for the discriminant function and for the training
data used to compute the discriminant function. Most statistics computed are
implementations of methods of the
discrim object. See the STRUCTURE section
above for the various implementations.
McLachlan, G. J. (1992).
Discriminant Analysis and Statistical Pattern Recognition,
John Wiley & Sons.
Seber, G.A.F. (1984).
Multivariate Observations,
John Wiley & Sons.
# Iris data
iris.quad <- discrim(Species~Sepal.L. + Sepal.W. + Petal.L. + Petal.W.,
data=iris.mm, family=Classical("heter"))
summary(iris.quad)