This function marks the features of an object of class
gbm.object.
Usage
mark_gbm(
object,
marking.system = c("presence", "influence", "normalized_influence")
)Arguments
- marking.system
a length-one character vector, indicating a strategy for marking the features. Available options are
"presence"(default),"influence", and"normalized_influence"- obejct
an object of class Trained
Details
The features in the Trained object in input
are marked, i.e. a numerical value is assigned to each feature
based on the choice of marking.system.
If marking.system = "presence", a binary value of 1/0 is assigned
to the features by their presence in the model:
1 is given to features that were used in the forest,
0 otherwise.
If marking.system = "influence", the relative influence of each
variable in reducing the loss function in the gbm object is used.
If marking.system = "normalized_influence", the relative influence
normalized to sum to 100 is used.