1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
|
% file MASS/predict.lda.d
% copyright (C) 1994-9 W. N. Venables and B. D. Ripley
%
\name{predict.lda}
\alias{predict.lda}
\title{
Classify Multivariate Observations by Linear Discrimination
}
\description{
Classify multivariate observations in conjunction with \code{lda}, and also
project data onto the linear discriminants.
}
\usage{
\method{predict}{lda}(object, newdata, prior = object$prior, dimen,
method = c("plug-in", "predictive", "debiased"), \dots)
}
\arguments{
\item{object}{
object of class \code{"lda"}
}
\item{newdata}{
data frame of cases to be classified or, if \code{object}
has a formula, a data frame with columns of the same names as the
variables used. A vector will be interpreted
as a row vector. If newdata is missing, an attempt will be
made to retrieve the data used to fit the \code{lda} object.
}
\item{prior}{
The prior probabilities of the classes, by default the proportions in the
training set or what was set in the call to \code{lda}.
}
\item{dimen}{
the dimension of the space to be used. If this is less than \code{min(p, ng-1)},
only the first \code{dimen} discriminant components are used (except for
\code{method="predictive"}), and only those dimensions are returned in \code{x}.
}
\item{method}{
This determines how the parameter estimation is handled. With \code{"plug-in"}
(the default) the usual unbiased parameter estimates are used and
assumed to be correct. With \code{"debiased"} an unbiased estimator of
the log posterior probabilities is used, and with \code{"predictive"} the
parameter estimates are integrated out using a vague prior.
}
\item{\dots}{
arguments based from or to other methods
}}
\value{
a list with components
\item{class}{
The MAP classification (a factor)
}
\item{posterior}{
posterior probabilities for the classes
}
\item{x}{
the scores of test cases on up to \code{dimen} discriminant variables
}}
\details{
This function is a method for the generic function \code{predict()} for
class \code{"lda"}. It can be invoked by calling \code{predict(x)} for
an object \code{x} of the appropriate class, or directly by calling
\code{predict.lda(x)} regardless of the class of the object.
Missing values in \code{newdata} are handled by returning \code{NA} if the
linear discriminants cannot be evaluated. If \code{newdata} is omitted and
the \code{na.action} of the fit omitted cases, these will be omitted on the
prediction.
This version centres the linear discriminants so that the
weighted mean (weighted by \code{prior}) of the group centroids is at
the origin.
}
\references{
Venables, W. N. and Ripley, B. D. (2002)
\emph{Modern Applied Statistics with S.} Fourth edition. Springer.
Ripley, B. D. (1996)
\emph{Pattern Recognition and Neural Networks}. Cambridge University Press.
}
\seealso{
\code{\link{lda}}, \code{\link{qda}}, \code{\link{predict.qda}}
}
\examples{
data(iris3)
tr <- sample(1:50, 25)
train <- rbind(iris3[tr,,1], iris3[tr,,2], iris3[tr,,3])
test <- rbind(iris3[-tr,,1], iris3[-tr,,2], iris3[-tr,,3])
cl <- factor(c(rep("s",25), rep("c",25), rep("v",25)))
z <- lda(train, cl)
predict(z, test)$class
}
\keyword{multivariate}
|
