McNemarTest<T>

Models & Types · AiDotNet.Evaluation.Statistics

McNemar’s test: compares the performance of two binary classifiers on the same dataset.

Implements: ITwoSampleTest<T>

For Beginners

McNemar’s test is specifically designed for comparing two classifiers’ predictions on the same dataset. It uses a 2x2 contingency table:

The test focuses on the disagreement cells (b and c) - cases where one model is right and the other is wrong.

How It Works

When to use:

  • Comparing two classifiers on the same test set
  • Binary classification problems
  • When you want to test if classifier A is significantly better than B

Constructors

Constructor Summary
McNemarTest()

Properties

Property Summary
string Description { get; }
string Name { get; }

Methods

Method Summary
StatisticalTestResult<T> Test(T[] correctA, T[] correctB, double alpha) Performs McNemar’s test using binary predictions.