-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathNIF-issue-2.ttl
24 lines (19 loc) · 968 Bytes
/
NIF-issue-2.ttl
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
# https://github.com/NLP2RDF/specification/issues/2
# itsrdf vs fise properties
</char=0,7> a nif:Word;
nif:anchorOf "Germany".
# If you have one NER tool and one annotation (not multiple candidates), NIF says use itsrdf properties.
</char=0,7>
itsrdf:taClassRef nerd:Country;
itsrdf:taIdentRef dbp:Germany;
itsrdf:taAnnotatorsRef "some tool";
itsrdf:confidence 0.9.
# If you have multiple annotations, you need to use fise properties (the NIF Stanbol profile), and node per annotation
</char=0,7/enrichment=1> a fise:EntityAnnotation;
fise:extracted-from <#char=0,7>;
fise:entity-type nerd:Country;
fise:entity-reference dbp:Germany;
dc:creator "some tool";
fise:confidence "0.9"^^xsd:float.
# I think it's a very bad practice to use two completely different property sets for these two situations.
# Just because in the second case there's an intermediate node for the annotation, doesn't mean the properties should be completely different.