predict: Prediction Method

In dtrSurv: Dynamic Treatment Regimes for Survival Analysis

Prediction Method

Description

Method to estimate the value for new data or to retrieve estimated value for training data

Usage

## S4 method for signature 'DTRSurv'
predict(object, ..., newdata, stage = 1, findOptimal = TRUE)

Arguments

object	A DTRSurv object. The object returned by a call to dtrSurv().
...	Ignored. Used to require named inputs.
newdata	NULL or a data.frame object. If NULL, this method retrieves the estimated value for the training data. If a data.frame, the value is estimated based on the data provided.
stage	An integer object. The stage for which predictions are desired.
findOptimal	A logical object. If TRUE, the value is estimated for all treatment options and that leading to the maximum value for each individual is used to estimate the value.

Value

a list object containing a matrix of the predicted survival function (survFunc), the estimated mean survuval (mean), and the estimated survival probability (if critical value is surv.mean or surv.prob)

Examples

dt <- data.frame("Y.1" = sample(1:100,100,TRUE), "Y.2" = sample(101:200,100,TRUE),
                 "D.1" = rbinom(100, 1, 0.9), "D.2" = rbinom(100,1,0.9),
                 "A.1" = rbinom(100, 1, 0.5), "A.2" = rbinom(100,1,0.5),
                 "X.1" = rnorm(100), "X.2" = rnorm(100))

result <- dtrSurv(data = dt, 
                  txName = c("A.1", "A.2"), 
                  models = list(Surv(Y.1,D.1)~X.1+A.1, 
                                Surv(Y.2,D.2)~X.2+A.2+Y.1))

tt <- predict(object = result)
tt <- predict(object = result, stage = 1)
tt <- predict(object = result, findOptimal = FALSE)
tt <- predict(object = result, newdata = dt)
tt <- predict(object = result, newdata = dt, stage = 1)
tt <- predict(object = result, newdaata = dt, findOptimal = FALSE)

dtrSurv documentation built on June 7, 2022, 5:07 p.m.