Description

References:

Implicitness of relationship grouping

Presentation at Wellington meeting October 2016

Presentation on progress

Objectives

• Clarify if the inconsistent application of role grouping is causing a problem, and if so, for whom.   Is there any clinical impact, for example?

  • Snorocket classifier has some assumption built in to avoid this problem, but other DL classifiers by default are apparently not able to give consistent results when dealing with relationships in more details.   (Reference?)

• Similarly, list the benefits (immediate and future) of 

• Generate formal guidance on use of Role Groups.

Impact

• Affects ECL in that requests which expect attributes to be grouped will not match ungrouped attributes.

Changes Required

List here changes that will need to be made to accommodate the proposal:

• Content changes to bring existing issues into line with guidance.

• Changes to the RF2→OWL conversion script.

• Authoring tooling: 

  • will need to accept single relationships in a non-zero group

• Validation changes:

  • Enhance rules around role groups

Sub Team

• @Daniel Karlsson (Unlicensed)

• @stefan.schulz (Unlicensed)

• @Former user (Deleted)

• @Yongsheng Gao

Child Pages 

Discussion

Relevant Documents