learn_decisions: Decision tree learning from modules

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

Decision tree learning from modules

Description

learn_decisions uses ranger to perform feature selection with respect to raw_modules.

Usage

learn_decisions(
  raw_modules,
  features,
  target,
  flatten.sep = "$",
  importance = "impurity",
  splitrule = "gini"
)

Arguments

raw_modules	list of numeric vectors. The raw modules.
features	numeric matrix or 3D array. The features to train on.
target	numeric vector. The target to train towards.
flatten.sep	string. Separator to use when flattening features.
importance	variable importance mode. See ranger:rangerranger::ranger.
splitrule	Splitting rule. See ranger:rangerranger::ranger.

Value

A list of shape (trees, modules, modules.weights), where modules are the sorted raw_modules with individual weights modules.weights, and trees contains one ranger decision tree per module.

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