Extract Node List from GBM Model — nodelist.gbm • networkformat

Extracts node-level attributes from a gbm model via pretty.gbm.tree(). Missing-sentinel nodes are excluded by keeping only nodes reachable through LeftNode/RightNode edges from the root.

Usage

# S3 method for class 'gbm'
nodelist(input_object, treenum = NULL, ...)

Arguments

input_object: A gbm model object from the gbm package
treenum: Integer vector of 1-based tree numbers to extract (default: NULL extracts all trees).
...: Additional arguments (currently unused)

Value

A data.frame with one row per real node and the following columns:

name: 0-based node ID within the tree
is_leaf: Logical: TRUE for terminal nodes
split_var: 0-based variable index (NA for leaves)
split_var_name: Variable name (NA for leaves)
split_point: Split threshold (NA for leaves)
prediction: Prediction value at the node
error_reduction: Impurity reduction from the split (0 for leaves); per-node variable importance from ErrorReduction
weight: Number of training observations reaching the node, from Weight
treenum: 1-based tree number
label: Display label: "<var>\n< <threshold>" for splits, rounded prediction for leaves

Examples

if (requireNamespace("gbm", quietly = TRUE)) {
  set.seed(1)
  fit <- gbm::gbm(mpg ~ ., data = mtcars,
                   distribution = "gaussian", n.trees = 5,
                   interaction.depth = 3, n.minobsinnode = 3)
  nl <- nodelist(fit)
  head(nl)
}
#>   name is_leaf split_var split_var_name split_point prediction error_reduction
#> 1    0   FALSE         0            cyl         5.0 -0.0290625      406.003333
#> 2    1    TRUE        NA           <NA>          NA  0.8434375        0.000000
#> 3    2   FALSE         2             hp       177.5 -0.3198958       68.728595
#> 4    3   FALSE         0            cyl         7.0 -0.1176339        6.518571
#> 5    4    TRUE        NA           <NA>          NA -0.0340625        0.000000
#> 6    5    TRUE        NA           <NA>          NA -0.2290625        0.000000
#>   weight treenum       label
#> 1     16       1    cyl\n< 5
#> 2      4       1      0.8434
#> 3     12       1 hp\n< 177.5
#> 4      7       1    cyl\n< 7
#> 5      4       1     -0.0341
#> 6      3       1     -0.2291