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
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/COV.R
\name{COV}
\alias{COV}
\alias{COV.ten}
\alias{COV.stratTen}
\alias{COV.numeric}
\title{\bold{cov}ariance matrix for survival data}
\usage{
COV(x, ...)
\method{COV}{ten}(x, ..., reCalc = FALSE)
\method{COV}{stratTen}(x, ..., reCalc = FALSE)
\method{COV}{numeric}(x, ..., n, ncg)
}
\arguments{
\item{x}{A \code{numeric} vector of
\emph{number of events}, \eqn{e_t}{e[t]}.
These are assumed to be ordered by discrete times.
\cr
A method is available for objects of \code{class} \code{ten}.}
\item{...}{Additional arguments (not implemented).}
\item{reCalc}{Recalcuate the values?
\cr
If \code{reCalc=FALSE} (the default) and the \code{ten} object already has
the calculated values stored as an \code{attribute},
the value of the \code{attribute} is returned directly.
\cr \cr
\bold{--Arguments for the numeric method:}}
\item{n}{\bold{n}umber at risk (total).}
\item{ncg}{\bold{n}umber at risk, per \bold{c}ovariate \bold{g}roup.
\cr
If there are \eqn{2} groups, this can be given as a \code{vector} with
the number at risk for group \eqn{1}.
\cr
If there are \eqn{\geq 2}{>= 2} groups, it is
a \code{matrix} with one column for each group.}
}
\value{
An \code{array}.
\cr
The first two dimensions = the number of covariate groups \eqn{K},
\eqn{k = 1, 2, \ldots K}.
This is the square matrix below.
\cr
The third dimension is the number of observations
(discrete time points).
\cr \cr
To calculate this, we use \code{x} (= \eqn{e_t}{e[t]} below) and
\eqn{n_1}{n1}, the number at risk in covariate group \eqn{1}.
\cr
Where there are \eqn{2} groups, the resulting sparse square matrix
(i.e. the non-diagonal elements are \eqn{0})
at time \eqn{t} has diagonal elements:
\deqn{cov_t = - \frac{n_{0t} n_{1t} e_t (n_t - e_t)}{n_t^2(n_t-1)}}{
cov[t] = - n0[t] * n1[t] * e[t] * (n[t] - e[t]) /
(n[t]^2 * (n[t] - 1))}
For \eqn{\geq 2}{>=2} groups, the resulting square matrix
has diagonal elements given by:
\deqn{cov_{kkt} = \frac{n_{kt}(n_t - n_{kt}) e_t(n_t - e_t)}{
n_t^2(n_t - 1)}}{
cov[k, k, t] = n[k, t] * (n[t] - n[k, t]) * e[t] * (n[t] - e[t]) /
(n[t]^2 * (n[t] - 1))}
The off diagonal elements are:
\deqn{cov_{klt} = \frac{-n_{kt} n_{lt} e_t (n_t-e_t) }{
n_t^2(n_t-1)}}{
cov[k, l, t] = - n[k, t] * n[l, t] * e[t] * (n[t] - e[t]) /
n[t]^2 * (n[t] - 1)}
}
\description{
\bold{cov}ariance matrix for survival data
}
\details{
Gives variance-covariance matrix for comparing survival
data for two or more groups.
\cr
Inputs are vectors corresponding to observations at a set of discrete
time points for right censored data, except for \eqn{n1},
the no. at risk by predictor.
\cr
This should be specified as a vector for one group,
otherwise as a matrix with each column corresponding to a group.
}
\note{
Where the is just one subject at risk \eqn{n=1} at
the final timepoint, the equations above may produce \code{NaN}
due to division by zero. This is converted to \code{0} for
simplicity.
}
\examples{
## Two covariate groups
## K&M. Example 7.2, pg 210, table 7.2 (last column).
\dontrun{
data("kidney", package="KMsurv")
k1 <- with(kidney,
ten(Surv(time=time, event=delta) ~ type))
COV(k1)[COV(k1) > 0]
}
## Four covariate groups
## K&M. Example 7.6, pg 217.
\dontrun{
data("larynx", package="KMsurv")
l1 <- ten(Surv(time, delta) ~ stage, data=larynx)
rowSums(COV(l1), dims=2)
}
## example of numeric method
## Three covariate groups
## K&M. Example 7.4, pg 212.
\dontrun{
data("bmt", package="KMsurv")
b1 <- asWide(ten(Surv(time=t2, event=d3) ~ group, data=bmt))
rowSums(b1[, COV(x=e, n=n, ncg=matrix(data=c(n_1, n_2, n_3), ncol=3))], dims=2)
}
}
\seealso{
Called by \code{\link{comp}}
The name of the function is capitalized
to distinguish it from:
\cr
?stats::cov
}
\keyword{survival}
|
