Importance of variables in a random forest

Source: R/measure_importance.R

Get a data frame with various measures of importance of variables in a random forest

measure_importance(forest, mean_sample = "top_trees", measures = NULL)

Arguments

forest

A random forest produced by the function randomForest with option localImp = TRUE
mean_sample

The sample of trees on which mean minimal depth is calculated, possible values are "all_trees", "top_trees", "relevant_trees"
measures

A vector of names of importance measures to be calculated - if equal to NULL then all are calculated; if "p_value" is to be calculated then "no_of_nodes" will be too. Suitable measures for classification forests are: mean_min_depth, accuracy_decrease, gini_decrease, no_of_nodes, times_a_root. For regression forests choose from: mean_min_depth, mse_increase, node_purity_increase, no_of_nodes, times_a_root.

Value

A data frame with rows corresponding to variables and columns to various measures of importance of variables

Examples

forest <- randomForest::randomForest(Species ~ ., data = iris, localImp = TRUE, ntree = 300)
measure_importance(forest)
#>       variable mean_min_depth no_of_nodes accuracy_decrease gini_decrease
#> 1 Petal.Length      1.1114865         749       0.298860983     42.010102
#> 2  Petal.Width      0.9932432         775       0.310185955     43.803248
#> 3 Sepal.Length      2.0128041         493       0.033657157     11.235496
#> 4  Sepal.Width      3.2915203         356       0.007638947      2.196532
#>   no_of_trees times_a_root      p_value
#> 1         296          116 3.594621e-13
#> 2         296          124 3.550723e-17
#> 3         265           60 9.999994e-01
#> 4         215            0 1.000000e+00