Continuous to Discrete Transformation
Discretizes continuous time variable into a specified grid of censored data for discrete survival analysis. It is a data preprocessing step, before the data can be extendend in long format and further analysed with discrete survival models.
contToDisc(dataSet, timeColumn, intervalLimits, equi=FALSE)
dataSet |
Original data in short format. Must be of class "data.frame". |
timeColumn |
Name of the column with discrete survival times. Must be a scalar character value. |
intervalLimits |
Numeric vector of the right interval borders, e. g. if the intervals are [0, a_1), [a_1, a_2), [a_2, a_max), then intervalLimits = c(a_1, a_2, a_max) |
equi |
Specifies if argument *intervalLimits* should be interpreted as number of equidistant intervals. Logical only TRUE or FALSE is allowed. |
Gives the data set expanded with a first column "timeDisc". This column includes the discrete time intervals (ordered factor).
In discrete survival analysis the survival times have to be categorized in time intervals. Therefore this function is required, if there are observed continuous survival times.
Thomas Welchowski welchow@imbie.meb.uni-bonn.de
Gerhard Tutz and Matthias Schmid, (2016), Modeling discrete time-to-event data, Springer series in statistics, Doi: 10.1007/978-3-319-28158-2
# Example copenhagen stroke study data
library(pec)
data(cost)
head(cost)
# Convert observed times to months
# Right borders of intervals [0, a_1), [a_1, a_2), ... , [a_{\max-1}, a_{\max})
IntBorders <- 1:ceiling(max(cost$time)/30)*30
# Select subsample
subCost <- cost [1:100, ]
CostMonths <- contToDisc (dataSet=subCost, timeColumn="time", intervalLimits=IntBorders)
head(CostMonths)
# Select subsample giving number of equidistant intervals
CostMonths <- contToDisc (dataSet=subCost, timeColumn="time", intervalLimits=10, equi=TRUE)
head(CostMonths)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.