Concept guide
Baseline
A declared reference method or case used to judge another result, with the comparison conditions made explicit.
Also used: reference method · comparator
Reading models and evidence
Eight useful concepts
Start with an example, then separate what a model answers from what it has yet to establish.
These are reader-facing explanations for the articles, rather than quotations of standards. Each entry keeps its example’s scope and reference sources visible.
Concept guide
Assumption
A condition taken as given when constructing a model or argument; its conclusions depend on the relevant conditions holding.
Also used: model assumption
Concept guide
Sensitivity
How a chosen output or conclusion changes when specified inputs or settings change. Global sensitivity analysis can attribute output uncertainty to uncertain inputs.
Also used: sensitivity analysis · numerical sensitivity
Concept guide
Identifiability
Whether the specified model, inputs and observations can distinguish parameter values. Structural identifiability asks about ideal information; practical identifiability asks what the actual finite, noisy experiment determines.
Also used: structural identifiability · practical identifiability
Concept guide
Calibration
For prediction intervals here, using held-out observed errors to adjust a band toward a stated coverage target.
Also used: interval calibration · prediction calibration
Concept guide
Validation
Assessing, with appropriate evidence, whether a model or procedure meets requirements for a specified use under stated conditions.
Also used: model validation · validation data
Concept guide
Robustness
The extent to which a specified result or level of performance survives specified changes in circumstances.
Also used: robustness bound
Concept guide
Uncertainty
Incomplete knowledge about a quantity, process or outcome. It may be represented by a probability model, a set of possibilities or a stated qualitative limitation.
Also used: measurement uncertainty · uncertainty set