clusterID
,
replicateID
,
recordID
.
recordDesign(cluster,data,record)
data
identifying the cluster variable or variables (in order).
Observations within a cluster may be correlated, and
records from different clusters are not correlated.
cluster
and
in
record
.
data
.
The data are grouped according to
cluster
,
and the ordering within clusters is determined by the order of the variable
names in
record
.
Variables in
record
may be either categorical or numeric,
although numeric variables are considered discrete
for the purposes of grouping.
If
record
is not specified,
data
is sorted by
cluster
, and the default
recordID
for a cluster is a vector of integers
from 1 to the size of the cluster.
"recordDesign"
, which is
a data frame grouped according to
cluster
and ordered within clusters as specified in
record
.
Three additional integer variables are appended to the data frame:
cluster
.record
.record
and (if there are replicates in a cluster) the original order of the data.cluster
and
record
) are appended to the data frame.
There are
"cluster"
and
"record"
attributes
which give the corresponding arguments to
recordDesign
.
This function is used for correlated data to create a mapping for cluster
and record identification.
Use
summary
to see a complete
cluster.
A warning is issued whenever
replicateID
and
recordID
are unequal,
indicating that the
record
specification
is not sufficient to produce
distinguishable records within clusters.
The variables in
cluster
must be distinct from variables in
record
.
Variables in
cluster
or
record
may be either categorical or numeric,
although numeric variables are considered discrete
for the purposes of grouping.
recordDesign(cluster="Dog",data=Pixel) recordDesign(cluster="Dog",data=Pixel,record="Side") recordDesign(cluster="Dog",data=Pixel,record="day") recordDesign(cluster="Dog",data=Pixel,record=c("Side","day")) recordDesign(cluster=c("Dog","Side"),data=Pixel,record="day")