Numerical approximation and error

On this page
  1. Mathematical statement
  2. Representative formula (exact TeX)
  3. Assumptions
  4. Symbol scope
  5. Intuition, definition, and misconceptions
  6. Adopted derivation guide
  7. Prerequisites
  8. Verification boundary
  9. Next obligation
  10. Same-source structured data

This English edition is an AI candidate and does not yet have an OCME Atlas (GPT-6) Twin closure receipt.

Absolute error is the absolute difference between approximation and reference; relative error additionally requires a nonzero reference and divides by its absolute value.

mko-numerical-approximation-error · definition · v0.10.0

Mathematical statement

For approximation a and reference x, absolute error is |a−x|. If x≠0, relative error is |a−x|/|x|. For tolerance τ≥0, being within tolerance means |a−x|≤τ.

Representative formula (exact TeX)

e=A(a,x)

Assumptions

  • Exercises use exact rationals rather than treating displayed binary floating-point values as exact answers.
  • Evaluate relative error only for a nonzero reference; truncation, rounding and data errors are not interchangeable.

Symbol scope

  • e — absolute error
  • a — approximation
  • x — reference value

Intuition, definition, and misconceptions

Absolute error retains the original quantity’s scale, whereas relative error compares it with a nonzero reference magnitude. They answer different questions.

The bound |a−x|≤τ is equivalent to a∈[x−τ,x+τ], including both endpoints.

An error value alone does not identify its cause. Distinguishing discretization, truncation, rounding, measurement and model errors requires information about how the values were produced.

Do not evaluate relative error at a zero reference or call every error a rounding error.

Adopted derivation guide

Numerical approximation and error: absolute error, relative error and tolerance for 0.33 approximating 1/3

The original MKO is retained. The guide has its own append-only revision and verification boundaries, not a new Lean proof.

Prerequisites

Verification boundary

Exact rational checks establish only the specified values and bounds; they assess neither algorithmic convergence nor a floating-point implementation.

not_formalized

This does not mean the content is false; it means no replayable computational or formal evidence is registered yet.

Next obligation

Add truncation/rounding sources, condition numbers, stability and convergence order.

Same-source structured data

This page and the interactive view use the same object and existing language edition; this adds no proof or human review.

Canonical MKO JSON · Plain Markdown