Criteria
- This panel provides information on every criterion used to subset detection & attribution methods when using the navigator.
- Criteria are grouped in five descriptive categories: Outcome, Data Compatibility, Assumptions, Model Properties, and Packages.
They capture:
- What a method aims to do
- Which data structures it requires or supports
- Under which foundational assumptions it remains valid
- How it behaves in terms of flexibility, scale, and feature support
- Where and how it can be implemented in practice
Criterion pages follow a common documentation format: Definition, Explanation, Tools/rationale for helping assessment and Example.
Common criterion options
Every criterion has the three base options to ease the assessment process: Any
, Don't know
and Inapplicable
. They are described below:
Option | Usage | Description |
---|---|---|
Any | Default when starting the criteria assessment | When selected, the criterion applies no filtering on the suggested method set. Users should consider every criterion of the chosen objective, and let the default only when all options are satisfying. |
Don't know | User don’t know which option to pick | When the user judges not being able to rightly assess the method against a criterion (lack of documentation, understanding, unsure), this option avoids subsetting methods by the corresponding criterion like the default Any . |
Inapplicable | Criterion cannot be applied because of incompatible paradigms | Because the navigator ambitions to be large and to include methods from different fields relying on heterogeneous assumptions, a criterion can be irrelevant to a case study. Methods indeed rely on implicit / explicit modelling paradigms that are not all focused on the same aspects and are not fully compatible with each other. |
Don't know
allows to reach the end of the assessment more easily but should be picked only when necessary since it will result in a larger and less informative set of suitable methods.? signs redirecting to the criterion page should allow the user to apply criteria in most cases.
When the criterion is understood but seems irrelevant to the question at hand,
Inapplicable
should be preferred.