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
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/bag.R
\docType{data}
\name{bag}
\alias{bag}
\alias{bag.default}
\alias{bagControl}
\alias{predict.bag}
\alias{ldaBag}
\alias{plsBag}
\alias{nbBag}
\alias{ctreeBag}
\alias{svmBag}
\alias{nnetBag}
\alias{print.bag}
\alias{summary.bag}
\alias{print.summary.bag}
\title{A General Framework For Bagging}
\format{
An object of class \code{list} of length 3.
An object of class \code{list} of length 3.
An object of class \code{list} of length 3.
An object of class \code{list} of length 3.
An object of class \code{list} of length 3.
An object of class \code{list} of length 3.
}
\usage{
bag(x, ...)
bagControl(
fit = NULL,
predict = NULL,
aggregate = NULL,
downSample = FALSE,
oob = TRUE,
allowParallel = TRUE
)
\method{bag}{default}(x, y, B = 10, vars = ncol(x), bagControl = NULL, ...)
\method{predict}{bag}(object, newdata = NULL, ...)
\method{print}{bag}(x, ...)
\method{summary}{bag}(object, ...)
\method{print}{summary.bag}(x, digits = max(3, getOption("digits") - 3), ...)
ldaBag
plsBag
nbBag
ctreeBag
svmBag
nnetBag
}
\arguments{
\item{x}{a matrix or data frame of predictors}
\item{\dots}{arguments to pass to the model function}
\item{fit}{a function that has arguments \code{x}, \code{y} and \code{...} and produces a model object #' that can later be used for prediction. Example functions are found in \code{ldaBag}, \code{plsBag}, #' \code{nbBag}, \code{svmBag} and \code{nnetBag}.}
\item{predict}{a function that generates predictions for each sub-model. The function should have #' arguments \code{object} and \code{x}. The output of the function can be any type of object (see the #' example below where posterior probabilities are generated. Example functions are found in \code{ldaBag}#' , \code{plsBag}, \code{nbBag}, \code{svmBag} and \code{nnetBag}.)}
\item{aggregate}{a function with arguments \code{x} and \code{type}. The function that takes the output #' of the \code{predict} function and reduces the bagged predictions to a single prediction per sample. #' the \code{type} argument can be used to switch between predicting classes or class probabilities for #' classification models. Example functions are found in \code{ldaBag}, \code{plsBag}, \code{nbBag}, #' \code{svmBag} and \code{nnetBag}.}
\item{downSample}{logical: for classification, should the data set be randomly sampled so that each #' class has the same number of samples as the smallest class?}
\item{oob}{logical: should out-of-bag statistics be computed and the predictions retained?}
\item{allowParallel}{a parallel backend is loaded and available, should the function use it?}
\item{y}{a vector of outcomes}
\item{B}{the number of bootstrap samples to train over.}
\item{vars}{an integer. If this argument is not \code{NULL}, a random sample of size \code{vars} is taken of the predictors in each bagging iteration. If \code{NULL}, all predictors are used.}
\item{bagControl}{a list of options.}
\item{object}{an object of class \code{bag}.}
\item{newdata}{a matrix or data frame of samples for prediction. Note that this argument must have a non-null value}
\item{digits}{minimal number of \emph{significant digits}.}
}
\value{
\code{bag} produces an object of class \code{bag} with elements
\item{fits }{a list with two sub-objects: the \code{fit} object has the actual model fit for that #' bagged samples and the \code{vars} object is either \code{NULL} or a vector of integers corresponding to which predictors were sampled for that model}
\item{control }{a mirror of the arguments passed into \code{bagControl}}
\item{call }{the call}
\item{B }{the number of bagging iterations}
\item{dims }{the dimensions of the training set}
}
\description{
\code{bag} provides a framework for bagging classification or regression models. The user can provide their own functions for model building, prediction and aggregation of predictions (see Details below).
}
\details{
The function is basically a framework where users can plug in any model in to assess
the effect of bagging. Examples functions can be found in \code{ldaBag}, \code{plsBag}
, \code{nbBag}, \code{svmBag} and \code{nnetBag}.
Each has elements \code{fit}, \code{pred} and \code{aggregate}.
One note: when \code{vars} is not \code{NULL}, the sub-setting occurs prior to the \code{fit} and #' \code{predict} functions are called. In this way, the user probably does not need to account for the #' change in predictors in their functions.
When using \code{bag} with \code{\link{train}}, classification models should use \code{type = "prob"} #' inside of the \code{predict} function so that \code{predict.train(object, newdata, type = "prob")} will #' work.
If a parallel backend is registered, the \pkg{foreach} package is used to train the models in parallel.
}
\examples{
## A simple example of bagging conditional inference regression trees:
data(BloodBrain)
## treebag <- bag(bbbDescr, logBBB, B = 10,
## bagControl = bagControl(fit = ctreeBag$fit,
## predict = ctreeBag$pred,
## aggregate = ctreeBag$aggregate))
## An example of pooling posterior probabilities to generate class predictions
data(mdrr)
## remove some zero variance predictors and linear dependencies
mdrrDescr <- mdrrDescr[, -nearZeroVar(mdrrDescr)]
mdrrDescr <- mdrrDescr[, -findCorrelation(cor(mdrrDescr), .95)]
## basicLDA <- train(mdrrDescr, mdrrClass, "lda")
## bagLDA2 <- train(mdrrDescr, mdrrClass,
## "bag",
## B = 10,
## bagControl = bagControl(fit = ldaBag$fit,
## predict = ldaBag$pred,
## aggregate = ldaBag$aggregate),
## tuneGrid = data.frame(vars = c((1:10)*10 , ncol(mdrrDescr))))
}
\author{
Max Kuhn
}
\keyword{datasets}
\keyword{models}
|
