Skip to content

Calculate Sorted Gini Importance

get_gini.Rd

Generate a table of sorted Gini importance scores for a random forest model.

Usage

get_gini(rf.model)

Arguments

rf.model

A random forest model object (see randomForest).

Value

A tibble of the sorted importance(s). MeanDecreaseAccuracy is the mean decrease in accuracy resulting from removing the feature.

See also

randomForest::randomForest()

Author

Stu Field

Examples

# Use tr_iris training data from iris data set
rf <- withr::with_seed(101, {
  randomForest::randomForest(
    Species ~ ., data = tr_iris, importance = TRUE,
    proximity = TRUE, keep.inbag = TRUE)
})
get_gini(rf)
#> # A tibble: 4 × 5
#>   Feature      Gini_Importance setosa virginica MeanDecreaseAccuracy
#>   <chr>                  <dbl>  <dbl>     <dbl>                <dbl>
#> 1 Sepal.Length           14.6  0.0585   0.0899               0.0726 
#> 2 Petal.Width            14.0  0.0382   0.0435               0.0395 
#> 3 Petal.Length           11.7  0.0365  -0.00261              0.0153 
#> 4 Sepal.Width             9.20 0.0107   0.00967              0.00982