Skip to contents

Accelerated ORSF control

Usage

orsf_control_fast(method = "efron", do_scale = TRUE, ...)

Arguments

method

(character) a character string specifying the method for tie handling. If there are no ties, all the methods are equivalent. Valid options are 'breslow' and 'efron'. The Efron approximation is the default because it is more accurate when dealing with tied event times and has similar computational efficiency compared to the Breslow method.

do_scale

(logical) if TRUE, values of predictors will be scaled prior to each instance of Newton Raphson scoring, using summary values from the data in the current node of the decision tree.

...

Further arguments passed to or from other methods (not currently used).

Value

an object of class 'orsf_control', which should be used as an input for the control argument of orsf.

Details

code from the survival package was modified to make this routine.

Adjust do_scale at your own risk. Setting do_scale = FALSE will reduce computation time but will also make the orsf model dependent on the scale of your data, which is why the default value is TRUE. It would be a good idea to center and scale your predictors prior to running orsf() if you plan on setting do_scale = FALSE.

See also

linear combination control functions orsf_control_cph(), orsf_control_custom(), orsf_control_net()

Examples


orsf(data = pbc_orsf,
     formula = Surv(time, status) ~ . - id,
     control = orsf_control_fast())
#> ---------- Oblique random survival forest
#> 
#>      Linear combinations: Accelerated
#>           N observations: 276
#>                 N events: 111
#>                  N trees: 500
#>       N predictors total: 17
#>    N predictors per node: 5
#>  Average leaves per tree: 25
#> Min observations in leaf: 5
#>       Min events in leaf: 1
#>           OOB stat value: 0.84
#>            OOB stat type: Harrell's C-statistic
#>      Variable importance: anova
#> 
#> -----------------------------------------