Finds the compositions and sizes of risk sets

Focus is on the risk set composition just prior to a failure.

risksets(
  x,
  strata = NULL,
  max.survs = NULL,
  members = TRUE,
  collate_sets = FALSE
)

Arguments

x	A `Surv` object.
strata	Stratum indicator.
max.survs	Maximum number of survivors in each risk set. If smaller than the 'natural number', survivors are sampled from the present ones. No sampling if missing.
members	If TRUE, all members of all risk sets are listed in the resulting list, see below.
collate_sets	logical. If TRUE, group information by risk sets in a list. Only if `members = TRUE`.

Value

A list with components (if collate_sets = FALSE)

antrs

No. of risk sets in each stratum. The number of strata is given by length(antrs).

risktimes

Ordered distinct failure time points.

eventset

If 'members' is TRUE, a vector of pointers to events in each risk set, else NULL.

riskset

If 'members' is TRUE, a vector of pointers to the members of the risk sets, in order. The 'n.events' first are the events. If 'members' is FALSE, 'riskset' is NULL.

size

The sizes of the risk sets.

n.events

The number of events in each risk set.

sample_fraction

If 'members' is TRUE, the sampling fraction of survivors in each risk set.

Details

If the input argument max.survs is left alone, all survivors are accounted for in all risk sets.

Note

Can be used to "sample the risk sets".

Author

Göran Broström

Examples


 enter <- c(0, 1, 0, 0)
 exit <- c(1, 2, 3, 4)
 event <- c(1, 1, 1, 0)
 risksets(Surv(enter, exit, event))