prune

Produce sequence of regression subtrees by pruning

Syntax

tree1 = prune(tree)

tree1 = prune(tree,Name=Value)

Description

tree1 = prune(tree) creates a copy of the regression tree tree with its optimal pruning sequence filled in.

tree1 = prune(tree,Name=Value) creates a pruned tree with additional options specified by a single name-value argument.

Examples

Prune Regression Tree

Load the carsmall data set. Consider Horsepower and Weight as predictor variables.

load carsmall;
X = [Weight Horsepower];
varNames = ["Weight" "Horsepower"];

Grow a regression tree using the entire data set. View the tree.

Mdl = fitrtree(X,MPG,PredictorNames=varNames)

Mdl = 
  RegressionTree
           PredictorNames: {'Weight'  'Horsepower'}
             ResponseName: 'Y'
    CategoricalPredictors: []
        ResponseTransform: 'none'
          NumObservations: 94

view(Mdl,Mode="graph");

The regression tree has 16 pruning levels.

Prune the regression tree to pruning-level 10. View the pruned tree.

MdlPruned = prune(Mdl,Level=10);
view(MdlPruned,Mode="graph");

The pruned tree has six pruning levels.

Input Arguments

tree — Regression tree

RegressionTree object

Regression tree, specified as a RegressionTree object created using the fitrtree function.

Name-Value Arguments

Specify optional pairs of arguments as Name1=Value1,...,NameN=ValueN, where Name is the argument name and Value is the corresponding value. Name-value arguments must appear after other arguments, but the order of the pairs does not matter.

Example: tree1 = prune(tree,Level=10) prunes the tree to level ten.

Before R2021a, use commas to separate each name and value, and enclose Name in quotes.

Example: tree1 = prune(tree,"Level",10) prunes the tree to level ten.

Alpha — Pruning cost

numeric scalar

Pruning cost, specified as a numeric scalar from 0 (no pruning) to 1 (prune to one node). The prune function prunes the tree to minimize the sum of (Alpha times the number of leaf nodes) and a cost (mean squared error).

Use only one of these three name-value arguments at a time: Alpha, Level, or Nodes.

Level — Pruning level

numeric scalar

Pruning level, specified as a numeric scalar from 0 (no pruning) to the largest pruning level of this tree max(tree.PruneList). The prune function returns the tree pruned to this level.

Use only one of these three name-value arguments at a time: Alpha, Level, or Nodes.

Nodes — Branch nodes to turn into leaf nodes

numeric vector

Branch nodes to turn into leaf nodes, specified as a numeric vector with elements from 1 to tree.NumNodes. Any tree branch nodes listed in Nodes become leaf nodes in tree1, unless their parent nodes are also pruned.

Use only one of these three name-value arguments at a time: Alpha, Level, or Nodes.

Output Arguments

tree1 — Pruned regression tree

RegressionTree object

Pruned regression tree, returned as a RegressionTree object.

Tips

• tree1 = prune(tree) returns the decision tree tree1 that is the full, unpruned tree, but with optimal pruning information added. This is useful only if you created tree by pruning another tree, or by using fitrtree with pruning set "off". If you plan to prune a tree multiple times along the optimal pruning sequence, it is more efficient to create the optimal pruning sequence first.

Extended Capabilities

GPU Arrays

Accelerate code by running on a graphics processing unit (GPU) using Parallel Computing Toolbox™.

Usage notes and limitations:

• prune does not execute pruning (estimation of the optimal sequence of pruned subtrees) on a GPU.

For more information, see Run MATLAB Functions on a GPU (Parallel Computing Toolbox).

Version History

Introduced in R2011a