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
|
\name{epi.cpresids}
\alias{epi.cpresids}
\title{
Covariate pattern residuals from a logistic regression model
}
\description{
Returns covariate pattern residuals and delta betas from a logistic regression model.
}
\usage{
epi.cpresids(obs, fit, covpattern)
}
\arguments{
\item{obs}{a vector of observed values (i.e., counts of `successes') for each covariate pattern).}
\item{fit}{a vector defining the predicted (i.e., fitted) probability of success for each covariate pattern.}
\item{covpattern}{a \code{\link{epi.cp}} object.}
}
\value{
A data frame with 13 elements: \code{cpid} the covariate pattern identifier, \code{n} the number of subjects in this covariate pattern, \code{obs} the observed number of successes, \code{pred} the predicted number of successes, \code{raw} the raw residuals, \code{sraw} the standardised raw residuals, \code{pearson} the Pearson residuals, \code{spearson} the standardised Pearson residuals, \code{deviance} the deviance residuals, \code{leverage} leverage, \code{deltabeta} the delta-betas, \code{sdeltabeta} the standardised delta-betas, and \code{deltachi} delta chi statistics.
}
\references{
Hosmer DW, Lemeshow S (1989). Applied Logistic Regression. John Wiley & Sons, New York, USA, pp. 137 - 138.
}
\seealso{
\code{\link{epi.cp}}
}
\examples{
## EXAMPLE 1:
dat.glm01 <- glm(case ~ spontaneous + induced, family = binomial(),
data = infert)
## Covariate patterns:
dat.mf01 <- model.frame(dat.glm01)
dat.cp01 <- epi.cp(dat.mf01[-1])
dat.obs01 <- as.vector(by(infert$case, as.factor(dat.cp01$id),
FUN = sum))
dat.fit01 <- as.vector(by(fitted(dat.glm01), as.factor(dat.cp01$id),
FUN = min))
dat.cpr01 <- epi.cpresids(obs = dat.obs01, fit = dat.fit01,
covpattern = dat.cp01)
head(dat.cpr01)
}
\keyword{univar}% at least one, from doc/KEYWORDS
\keyword{univar}% __ONLY ONE__ keyword per line
|
