RUS ranger

Source: R/rusranger.R

Adapted default settings to the ranger() to support random-under-sampling. Additionally the default settings are modified to the most common classification settings used in the AMPEL project.

rusranger(
  x,
  y,
  probability = TRUE,
  classification = !probability,
  min.node.size = if (probability) 10 else 1,
  replace = FALSE,
  ...
)

Arguments

x

matrix/data.frame, feature matrix, see ranger() for details.

y

numeric/factor, classification labels, see ranger() for details.

probability

logical(1), grow probability trees, see ranger() for details.

classification

logical(1), run classification even if y is numeric, see ranger() for details.

min.node.size,

same as in ranger()

replace,

subsampling without (default, replace = FALSE) or with resampling, see ranger() for details.

...

further arguments passed to ranger().

Value

ranger object, see ranger() for details.

Details

In contrast to ranger() rusranger() currently just supports binary classifications.

References

AMPEL project: Analysis and Reporting System for the Improvement of Patient Safety through Real-Time Integration of Laboratory Findings, https://ampel.care.

See also

ranger()