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
|
\name{add_err}
\alias{add_err}
\title{
Add normally distributed errors to simulated kinetic degradation data
}
\description{
Normally distributed errors are added to data predicted for a specific
degradation model using \code{\link{mkinpredict}}. The variance of the error
may depend on the predicted value and is specified as a standard deviation.
}
\usage{
add_err(prediction, sdfunc, secondary = c("M1", "M2"),
n = 1000, LOD = 0.1, reps = 2,
digits = 1, seed = NA)
}
\arguments{
\item{prediction}{
A prediction from a kinetic model as produced by \code{\link{mkinpredict}}.
}
\item{sdfunc}{
A function taking the predicted value as its only argument and returning
a standard deviation that should be used for generating the random error
terms for this value.
}
\item{secondary}{
The names of state variables that should have an initial value of zero
}
\item{n}{
The number of datasets to be generated.
}
\item{LOD}{
The limit of detection (LOD). Values that are below the LOD after adding
the random error will be set to NA.
}
\item{reps}{
The number of replicates to be generated within the datasets.
}
\item{digits}{
The number of digits to which the values will be rounded.
}
\item{seed}{
The seed used for the generation of random numbers. If NA, the seed
is not set.
}
}
\value{
A list of datasets compatible with \code{\link{mmkin}}, i.e.
the components of the list are datasets compatible with
\code{\link{mkinfit}}.
}
\references{
Ranke J and Lehmann R (2015) To t-test or not to t-test, that is the question. XV Symposium on Pesticide Chemistry 2-4 September 2015, Piacenza, Italy
http://chem.uft.uni-bremen.de/ranke/posters/piacenza_2015.pdf
}
\author{
Johannes Ranke
}
\examples{
# The kinetic model
m_SFO_SFO <- mkinmod(parent = mkinsub("SFO", "M1"),
M1 = mkinsub("SFO"), use_of_ff = "max")
# Generate a prediction for a specific set of parameters
sampling_times = c(0, 1, 3, 7, 14, 28, 60, 90, 120)
# This is the prediction used for the "Type 2 datasets" on the Piacenza poster
# from 2015
d_SFO_SFO <- mkinpredict(m_SFO_SFO,
c(k_parent = 0.1, f_parent_to_M1 = 0.5,
k_M1 = log(2)/1000),
c(parent = 100, M1 = 0),
sampling_times)
# Add an error term with a constant (independent of the value) standard deviation
# of 10, and generate three datasets
d_SFO_SFO_err <- add_err(d_SFO_SFO, function(x) 10, n = 3, seed = 123456789 )
# Name the datasets for nicer plotting
names(d_SFO_SFO_err) <- paste("Dataset", 1:3)
# Name the model in the list of models (with only one member in this case)
# for nicer plotting later on.
# Be quiet and use the faster Levenberg-Marquardt algorithm, as the datasets
# are easy and examples are run often. Use only one core not to offend CRAN
# checks
f_SFO_SFO <- mmkin(list("SFO-SFO" = m_SFO_SFO),
d_SFO_SFO_err, cores = 1,
quiet = TRUE, method.modFit = "Marq")
plot(f_SFO_SFO)
# We would like to inspect the fit for dataset 3 more closely
# Using double brackets makes the returned object an mkinfit object
# instead of a list of mkinfit objects, so plot.mkinfit is used
plot(f_SFO_SFO[[3]], show_residuals = TRUE)
# If we use single brackets, we should give two indices (model and dataset),
# and plot.mmkin is used
plot(f_SFO_SFO[1, 3])
}
\keyword{ manip }
|
