predict_cox_reg: Compute the prediction of a model built with 'cox_reg'
In YohannLeFaou/sword: Random forest with IPC weights for survival analysis

Description Usage Arguments Value See Also Examples

Given a model built wtih cox_reg, predict_cox_reg allows to get the predictions of the model for new observations

1	predict_cox_reg(obj, newdata)

`obj`	A list output by `cox_reg`
`newdata`	A data.frame which contains the same variables as the ones used for the training

A list with the following elements :

`pred`	The vector of the predicted values for `phi`(T') (with T' = min(T, `max_time`)) for the observations of `newdata`
`surv`	The matrix which contains the estimated values of the survival curves at `time_points`, for the observations of `newdata`
`time_points`	The vector of the time points where the survival curves are evaluated

cox_reg

data("transplant", package = "survival")
transplant$delta = 1 * (transplant$event == "ltx") # create binary var
# which indicate censoring/non censoring

# keep only rows with no missing value
transplant_bis = transplant[stats::complete.cases(transplant),]


# ------------------------------------------------
#   Basic call to train a model
# ------------------------------------------------

set.seed(17)
train_lines = sample(1:nrow(transplant_bis), 600)
res1 = cox_reg(y_var = "futime",
                      delta_var = "delta",
                      x_vars = setdiff(colnames(transplant_bis),c("futime", "delta", "event")),
                      train = transplant_bis[train_lines,],
                      types_w_ev = c("KM", "Cox", "RSF", "unif"))

# ------------------------------------------------
#   Predict on new data
# ------------------------------------------------

pred1 = predict_cox_reg(obj = res1,
                               newdata = transplant_bis[-train_lines,])
print(pred1$pred[1:30])