make.time.factor: Make time-varying dummy variables from time-varying factor variable
Description
Create a new dataframe with time-varying dummy variables from a time-varying
factor variable. The time-varying dummy variables are named appropriately
to be used as a set of time dependent individual covariates in a parameter
specification
dataframe containing set of factor variables with names composed of
var.name prefix and times suffix
var.name
prefix for variable names
times
numeric suffixes for variable names
intercept
the value of the factor variable that will be used for the
intercept
delete
if TRUE, the origninal time-varying factor variables are
removed from the returned dataframe
Value
x: a dataframe containing the original data (with time-varying
factor variables removed if delete=TRUE) and the time-varying dummy
variables added.
Details
An example of the var.name and times is var.name="observer", times=1:5. The
code expects to find observer1,...,observer5 to be factor variables in x. If
there are k unique levels (excluding ".") across the time varying factor
variables, then k-1 dummy variables are created for each of the named factor
variables. They are named with var.name, level[i], times[j] concatenated
together where level[i] is the name of the facto level i. If there a m
times then the new data set will contain m*(k-1) dummy variables. If the
factor variable includes any "." values these are ignored because they are
used to indicate a missing value that is paired with a missing value in the
encounter history. Note that it will create each dummy variable for each
factor even if a particular level is not contained within a factor (eg
observers 1 to 3 used but only 1 and 2 on occasion 1).