Attendees
- Mandy Chessell
- Chris Grote (Deactivated)
- Chris Replogle
- David Radley
- Hibatallah Kabbaj
- Maryna Strelchuk.Strelchuk@ing.com
- Nigel Jones
- Rihi Ayoub
- William Bittles
- John Mertic
- Liz McIntosh
- Graham Wallis (Deactivated)
Agenda
- Review of the Analytics Model open metadata types
- Background models:
View file name analytical-models.drawio height 250
...
Notes and Review Documents
- Review of the Analytics Model open metadata types
- Agreed to change the proposed classification name for a
Project
fromAnalyticsModelProject
toAnalyticProject
to ensure that the project can be used for all things related to the analytic project (rather than suggesting it is only the analytical model). - Discussed the need to have the various
SoftwareComponent
s that are potentially re-used / re-usable refer to a specific version of theSoftwareComponent
they are re-using. (So that if theSoftwareComponent
they are re-using continues to evolve, we know precisely which version we are re-using.) - Discussed the
Variable
area, related to experiments, and concluded that this is a more general description of things like features (inputs) and outputs from analytical models – open questions around whether we need to (and how best to) clearly distinguish between the inputs and outputs (predictions) from models (e.g. like we have forPort
s andProcess
es, or using sub-types, classifications, or other approaches?) - Agreed it makes sense to introduce a new top-level type from which everything extends, so that we can have the
ExternalId
entity related to it even for objects that do not themselves extendReferenceable
(likeVariable
s,Like
s, etc)
- Agreed to change the proposed classification name for a