jointplot {joineR} | R Documentation |
This function views the longitudinal profile of each unit with event-time (censored or not) taken as the end-point, referred to as time zero. In doing so, the shape of the profile prior to event-time can be inspected. This can be done over a user-specified number of time units.
jointplot(longdat, survdat, lag = 5, split = TRUE)
longdat |
a matrix containing both the longitudinal data and the longitudinal design matrix. The matrix will contain one row per measurement per subject. The matrix is required to have the column form: subject index, longitudinal response, longitudinal measurement time, longitudinal design matrix. |
survdat |
matrix containing both the survival data and the survival design matrix and must contain one row per subject. The matrix is required to be set up in the following column order: subject index, survival time, censoring indicator, survival design matrix. |
lag |
Argument which specifies how many units in time we look back through. |
split |
TRUE/FALSE argument which allows the profiles of units which 'fail' and those which are 'censored' to be viewed in separate panels of the same graph. This is the default option. Using split=FALSE will plot all profiles overlaid on a single plot. |
The function tailors the 'xyplot' function in the 'lattice' library to produce a representation of joint data with longitudinal and survival components.
Pete Philipson (p.m.philipson@ncl.ac.uk)
Wulfsohn, M. S. and Tsiatis, A. A. (1997) "A Joint Model for Survival and Longitudinal Data Measured with Error", Biometrics, 53, 330-339.
'lattice','xyplot','joint','jointsep','jointcp'.
data(liverlong,liversurv) jointplot(liverlong,liversurv,lag=8)