Predictions from proportional hazards model
# S3 method for phreg
predict(object, newdata, times = NULL,
individual.time = FALSE, tminus = FALSE, se = TRUE,
robust = FALSE, conf.type = "log", conf.int = 0.95, ...)phreg object
data.frame
Time where to predict variable, default is all time-points from the object sorted
when TRUE then newdata and times have same length and makes only predictions for these individual times.
to make predictions in T- that is just before, useful for IPCW techniques
with standard errors and upper and lower confidence intervals.
to get robust se's.
transformation for suvival estimates, default is log
significance level
Additional arguments to plot functions