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
|
\name{regSim}
\alias{regSim}
\alias{LM3}
\alias{LOGIT3}
\alias{GAM3}
\title{Regression Model Simulation}
\description{
Simulates regression models.
}
\usage{
regSim(model = "LM3", n = 100, ...)
LM3(n = 100, seed = 4711)
LOGIT3(n = 100, seed = 4711)
GAM3(n = 100, seed = 4711)
}
\arguments{
\item{model}{
a character string defining the function name from which the
regression model will be simulated.
}
\item{n}{
an integer value setting the length, i.e. the number of records
of the output series, an integer value. By default \code{n=100}.
}
\item{seed}{
an integer value, the recommended way to specify seeds for
random number generation.
}
\item{\dots}{
arguments to be passed to the underlying function specified by
the \code{model} argument.
}
}
\details{
The function \code{regSim} allows to simulate from various regression
models defined by one of the three example functions \code{LM3},
\code{LOGIT3}, \code{GAM3} or by a user specified function.
The examples are defined in the following way:
\code{# LM3:}\cr
\code{> y = 0.75 * x1 + 0.25 * x2 - 0.5 * x3 + 0.1 * eps }\cr
\code{# LOGIT3:}\cr
\code{> y = 1 / (1 + exp(- 0.75 * x1 + 0.25 * x2 - 0.5 * x3 + eps)) }\cr
\code{# GAM3:}\cr
\code{> y = scale(scale(sin(2 * pi * x1)) + scale(exp(x2)) + scale(x3)) }\cr
\code{> y = y + 0.1 * rnorm(n, sd = sd(y))}\cr
\code{"LM3"} models a liner regression model, \code{"LOGIT3"} a generalized
linear regression model expressed by a logit model, and \code{"GAM"} an
additive model. \code{x1}, \code{x2}, \code{x3}, and \code{eps} are random
normal deviates of length \code{n}.
The \code{model} function should return an rectangular series defined
as an object of class \code{data.frame}, \code{timeSeries} or \code{mts}
which can be accepted from the parameter estimation
functions \code{regFit} and \code{gregFit}.
}
\value{
The function \code{garchSim} returns an object of the same class
as returned by the underlying function \code{match.fun(model)}.
These may be objects of class \code{data.frame}, \code{timeSeries} or
\code{mts}.
}
\note{
This function is still under development. For the future we plan,
that the function \code{regSim} will be able to generate general
regression models.
}
\author{
Diethelm Wuertz for the Rmetrics \R-port.
}
\examples{
## LM2 -
# Data for a user defined linear regression model:
LM2 = function(n){
x = rnorm(n)
y = rnorm(n)
eps = 0.1 * rnorm(n)
z = 0.5 + 0.75 * x + 0.25 * y + eps
data.frame(Z = z, X = x, Y = y)
}
for (FUN in c("LM2", "LM3")) {
cat(FUN, ":\n", sep = "")
print(regSim(model = FUN, n = 10))
}
}
\keyword{models}
|
