unique_module_importance: Calculate the importance of each unique module.

In pievos101/DFNET: Network-guided greedy decision forest for feature subset selection

Calculate the importance of each unique module.

Description

Calculate the importance of each unique module.

Usage

unique_module_importance(
  graph,
  modules,
  edge_importances,
  tree_importances,
  mc.cores = 1,
  collapse = mean
)

Arguments

graph	the graph used for training
modules	the modules obtained by training
edge_importances	the importance scores assigned to each edge in the graph, e.g. by edge_importance.
tree_importances	the importance scores assigned to each decision tree in the forest
mc.cores	how many cores to use in parallel
collapse	how accumulate multiple values per module

Value

the collapsed tree and edge importances of each unique module, as well as their sum.

See Also

Other metrics: edge_importance(), feature_importance(), module_importance()

pievos101/DFNET documentation built on Dec. 1, 2022, 3:44 p.m.