liveBook · Manning

Out-of-bag error

As explained in Timeless Algorithms, The Foundational Ideas

Out-of-bag error estimates prediction error using observations omitted from the bootstrap samples used to train the relevant trees.

Out-of-bag error estimates prediction error from observations omitted from the bootstrap samples used to train the relevant trees. section 9.4.2

It helps compare random-forest configurations as the number of candidate features at each split changes. For classification, compare it alongside strength and margin distributions; for regression, use out-of-bag mean squared error or mean absolute error. section 9.4.2

How Timeless Algorithms, The Foundational Ideas covers it

How deep it goes
Defines error estimation from omitted bootstrap observations and identifies comparison measures, but does not detail the calculation.
Examples
Compare classification forests with out-of-bag error and margins, or regression forests with mean squared or mean absolute error.
What you can do after
Evaluate forest configurations by comparing out-of-bag error and related measures.
How it connects
Connects out-of-bag error to random forests, tree strength, inter-tree correlation, and margin distributions.

Also explained in

Compare how 3 books explain out-of-bag error →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Timeless Algorithms, The Foundational Ideas. The section links point to the parts of the book it draws on.