In case, you are confused my semantic type groups are slightly different than these:
Over time, we have modified our approach in:
– assigning types to groups
– assigning labels to groups
– excluding types
For me, the object lesson has been to not assign types to groups too far upstream. And, as Jon or Brian said on the last ontology call, let the VIVO application handle the assignment of types to groups.
Janos:Jon and Paul,
I would think we want to define any new properties in terms of SKOS vocabulary to be consistent with the ISF and how we are handling external vocabulary imports. I would think we would have two vocabulary schemes one for UMLS CUIs and a secondary scheme for the semantic types.
cui:4567 skos:semanticRelation semantictype:4456 .
More detail can be found here:
Examples of SKOS vocabularies in the ISF can be found here http://code.google.com/p/connect-isf/source/browse/#svn%2Ftrunk%2Fsrc%2Fontology%2Fsource-skos
I hope I have not muddied the waters too much. I can see that this approach of categorizing concepts would be useful in the VIVO application.