Skip to contents

Calculate feature importance in terms of xgboost metrics and shap values

getFeatureImp.Rd

Calculate feature importance in terms of xgboost metrics and shap values

Usage

getFeatureImp(model, data, outcome)

Arguments

model

output xgboost model from `buildXGboost()`

data

dataframe, input data used to train the model

outcome

string, name of the outcome variable

Value

a data frame that contains 4 metrics for each feature

  • Gain: improvement in accuracy brought by the feature to the branches it is on

  • Cover: relative quantity of observations concerned by a feature

  • Frequency: the number of times a feature is used in all generated trees

  • shap: average Shapley value for the feature

Examples

if (FALSE) { # \dontrun{
getFeatureImp(xgb, voteData, "vote")
} # }