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
|
\name{logLik.mkinfit}
\alias{logLik.mkinfit}
\title{
Calculated the log-likelihood of a fitted mkinfit object
}
\description{
This function simply calculates the product of the likelihood densities
calculated using \code{\link{dnorm}}, i.e. assuming normal distribution.
The total number of estimated parameters returned with the value
of the likelihood is calculated as the sum of fitted degradation
model parameters and the fitted error model parameters.
For the case of unweighted least squares fitting, we calculate one
constant standard deviation from the residuals using \code{\link{sd}}
and add one to the number of fitted degradation model parameters.
For the case of manual weighting, we use the weight given for each
observation as standard deviation in calculating its likelihood
and the total number of estimated parameters is equal to the
number of fitted degradation model parameters.
In the case of iterative reweighting, the variances obtained by this
procedure are used in the likelihood calculations, and the number of
estimated parameters is obtained by the number of degradation model
parameters plus the number of variance model parameters, i.e. the number of
observed variables if the reweighting method is "obs", and two if the
reweighting method is "tc".
}
\usage{
\method{logLik}{mkinfit}(object, ...)
}
\arguments{
\item{object}{
An object of class \code{\link{mkinfit}}.
}
\item{\dots}{
For compatibility with the generic method
}
}
\value{
An object of class \code{\link{logLik}} with the number of
estimated parameters (degradation model parameters plus variance
model parameters) as attribute.
}
\seealso{
Compare the AIC of columns of \code{\link{mmkin}} objects using
\code{\link{AIC.mmkin}}.
}
\examples{
sfo_sfo <- mkinmod(
parent = mkinsub("SFO", to = "m1"),
m1 = mkinsub("SFO")
)
d_t <- FOCUS_2006_D
d_t[23:24, "value"] <- c(NA, NA) # can't cope with zero values at the moment
f_nw <- mkinfit(sfo_sfo, d_t, quiet = TRUE) # no weighting (weights are unity)
f_obs <- mkinfit(sfo_sfo, d_t, reweight.method = "obs", quiet = TRUE)
f_tc <- mkinfit(sfo_sfo, d_t, reweight.method = "tc", quiet = TRUE)
d_t$err <- d_t$value # Manual weighting assuming sigma ~ y
f_man <- mkinfit(sfo_sfo, d_t, err = "err", quiet = TRUE)
AIC(f_nw, f_obs, f_tc, f_man)
}
\author{
Johannes Ranke
}
|
