liveBook · Manning

Permutation test

As explained in Causal AI

A permutation test constructs a null distribution by repeatedly rearranging observed data or labels instead of relying on a canonical parametric test distribution.

A permutation test constructs a p-value by repeatedly rearranging observed data or labels rather than comparing a statistic with a canonical distribution. In conditional-independence testing, it offers an alternative when continuous-variable tests rely on linear-dependence assumptions. section 4.3.2

Use permutation tests when:

  • A canonical test distribution is unsuitable or its assumptions are too strong. section 4.3.2
  • Conditional-independence testing involves continuous variables where dependence may not be linear. section 4.3.2
  • Fewer assumptions justify accepting greater computational expense. section 4.3.2

How Causal AI covers it

How deep it goes
Defines the algorithmic rationale and assumption trade-off but does not detail permutation steps or implementation.
Examples
Uses continuous-variable conditional-independence tests as the motivating case, contrasting linear assumptions with permutation testing.
What it assumes
Assumes familiarity with conditional-independence tests, test statistics, p-values, and canonical test distributions.
What you can do after
Evaluate conditional independence with fewer distributional assumptions while accounting for higher computational cost.
How it connects
Connects permutation tests to conditional independence, test statistics, canonical distributions, and linear dependence.
Cover of Causal AI
liveBookmanning.com
Read Causal AI on liveBook Get the book at manning.com

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Causal AI. The section links point to the parts of the book it draws on.