capLargeValues {mlr} | R Documentation |
Convert large/infinite numeric values in a data.frame or task.
Description
Convert numeric entries which large/infinite (absolute) values
in a data.frame or task.
Only numeric/integer columns are affected.
Usage
capLargeValues(
obj,
target = character(0L),
cols = NULL,
threshold = Inf,
impute = threshold,
what = "abs"
)
Arguments
obj |
(data.frame | Task)
Input data.
|
target |
(character)
Name of the column(s) specifying the response.
Target columns will not be capped.
Default is character(0) .
|
cols |
(character)
Which columns to convert.
Default is all numeric columns.
|
threshold |
(numeric(1) )
Threshold for capping.
Every entry whose absolute value is equal or larger is converted.
Default is Inf .
|
impute |
(numeric(1) )
Replacement value for large entries.
Large negative entries are converted to -impute .
Default is threshold .
|
what |
(character(1) )
What kind of entries are affected?
“abs” means abs(x) > threshold ,
“pos” means abs(x) > threshold && x > 0 ,
“neg” means abs(x) > threshold && x < 0 .
Default is “abs”.
|
Value
(data.frame)
See Also
Other eda_and_preprocess:
createDummyFeatures()
,
dropFeatures()
,
mergeSmallFactorLevels()
,
normalizeFeatures()
,
removeConstantFeatures()
,
summarizeColumns()
,
summarizeLevels()
Examples
capLargeValues(iris, threshold = 5, impute = 5)
[Package
mlr version 2.19.2
Index]