liveBook · Manning

Margin

In classification, a margin is a measure of separation between a decision boundary and data points, or of the correct class’s score advantage over competing classes.

How 3 Manning books explain it

Timeless Algorithms, The Foundational Ideas

A boosting margin measures the combined classifier's support for the correct class relative to a competing class.

How it is used
Uses boosting margins to explain how an ensemble can improve on new data after reaching perfect training accuracy.
How deep it goes
Defines a boosting margin qualitatively and relates it to generalization, without specifying its calculation or a training procedure.
Examples
Contrasts random forest vote margins of 0.70 and 0.02 for two correct classifications.
What you can do after
Evaluate a classifier’s separation beyond accuracy and distinguish vote-based margins from an SVM’s geometric margin.

Read how Timeless Algorithms, The Foundational Ideas explains margin →

Machine Learning System Design

A measure of the separation between observations and the classifier boundary that determines their predicted classes.

How it is used
Uses metric sensitivity to sample noise to assess whether a classifier’s decision boundary is robust.
How deep it goes
Explains threshold sensitivity, noise testing, bootstrap distributions, and confidence bounds without deriving classifier geometry.
Examples
Demonstrates robustness using precision-recall distributions, bootstrapped datasets, sample sizes, and added normally distributed noise.
What you can do after
Diagnose boundary fragility by comparing noisy and unnoised metric distributions and reporting bootstrapped lower confidence bounds.

Read how Machine Learning System Design explains margin →

Machine Learning Algorithms in Depth

A margin is the distance between a classifier's decision boundary and the nearest training examples or class-separating boundaries.

How it is used
Uses geometric separation to select among correctly classifying decision boundaries in support vector machines.
How deep it goes
Derives a max-min distance objective and its quadratic-program formulation, then stops short of further optimization details.
Examples
Uses training examples, margin boundaries, support vectors, and a maximum-margin SVM diagram.
What you can do after
Evaluate classifiers by their nearest-example distance and understand why support vectors determine the maximum-margin solution.

Read how Machine Learning Algorithms in Depth explains margin →

Also covered in

What to know first

Related concepts

Questions about margin

What is margin?

In classification, a margin is a measure of separation between a decision boundary and data points, or of the correct class’s score advantage over competing classes.

Which Manning books explain margin?

Timeless Algorithms, The Foundational Ideas, Machine Learning System Design and Machine Learning Algorithms in Depth.

What should I know before learning margin?

Manning books that teach margin usually introduce convex optimization, decision boundary and classifier first.

What is the difference between margin and vote gap?

In classification, a margin is a measure of separation between a decision boundary and data points, or of the correct class’s score advantage over competing classes. A vote gap is the difference between the vote shares of an observation’s two leading predicted classes.

What is margin part of?

Margin is part of high-dimensional data. High-dimensional data consists of observations represented by a large number of features, variables, coordinates, or other dimensions.

These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.