Predict from a fitted survival learner

Produces predictions from a fit() object. Every matrix-valued type returns a survmat (an n x length(times) numeric matrix carrying its grid in the "times" attribute); the scalar-valued types ("rmst", "quantile", "median") return a numeric vector (or an n x length(probs) matrix for a vector probs).

Usage

# S3 method for class 'survalis_fit'
predict(
  object,
  newdata = NULL,
  times = NULL,
  type = c("survival", "risk", "chf", "hazard", "rmst", "quantile", "median"),
  ...
)

Arguments

object

A survalis_fit from fit().

newdata

A data frame of predictors. Defaults to the training data. Rows with missing values in model variables are rejected.

times

Numeric evaluation grid. When NULL, up to 50 event-time quantiles between the 5th and 95th percentiles are used and recorded on the result.

type

One of:

"survival"

predicted survival probability \(S(t)\) (default).

"risk"

cumulative risk \(1 - S(t)\).

"chf"

cumulative hazard \(-\log S(t)\).

"hazard"

piecewise-constant hazard on the grid.

"rmst"

restricted mean survival time; upper limit tau (...), default max(times).

"quantile"

survival-time quantile(s); requires probs (...).

"median"

median survival time (probs = 0.5).

...

Type-specific arguments: tau for "rmst", probs for "quantile".

Value

A survmat for "survival", "risk", "chf", "hazard"; a numeric vector for "rmst", "median" and scalar-probs "quantile"; an n x length(probs) matrix for vector probs.

Examples

m <- fit(Surv(time, status) ~ age + karno, veteran, "ranger",
         spec = list(num.trees = 100))
predict(m, veteran[1:5, ], times = c(90, 180, 365))
#> <survmat> survival  |  5 obs x 3 times  [90, 365]
#>           t=90     t=180     t=365
#> [1,] 0.2624219 0.1967211 0.1828112
#> [2,] 0.8763340 0.4313011 0.3130361
#> [3,] 0.7686504 0.4792258 0.2617366
#> [4,] 0.7931211 0.3105448 0.1516631
#> [5,] 0.8492272 0.3460189 0.2651582
predict(m, veteran[1:5, ], times = c(90, 180, 365), type = "risk")
predict(m, veteran[1:5, ], type = "rmst", tau = 365)
#>        1        2        3        4        5 
#> 119.4606 212.7497 181.8384 171.1826 188.9202