Discussion this morning with @AminaANNANE and @zevio (cc also @twktheainur ) about the idea of manually assigning semantic similarity within the STY in order to use such similarities to disambiguate annotations with multiples TUIs (and the GUIs).
The idea would consist in looking around an ambiguous annotation to see which pattern to prefer.
For instance,
"Traitement de la douleur des pancréatites chroniques par la neurotomie de la tête du pancréas ; une technique nouvelle et ses résultats . "
Generates an ambiguous annotation on "tête du pancréas"
Looking at the sentence with TUI will gives (annotated with SNMIFRE, CIM-10, MSHFRE:
T061 T184 T061 T023
or
T061 T184 T061 T191
Is (T061 T184 T061) "more semantically similar" to T023 or than T191 ?
Discussion this morning with @AminaANNANE and @zevio (cc also @twktheainur ) about the idea of manually assigning semantic similarity within the STY in order to use such similarities to disambiguate annotations with multiples TUIs (and the GUIs).
The idea would consist in looking around an ambiguous annotation to see which pattern to prefer.
For instance, "Traitement de la douleur des pancréatites chroniques par la neurotomie de la tête du pancréas ; une technique nouvelle et ses résultats . " Generates an ambiguous annotation on "tête du pancréas"
Looking at the sentence with TUI will gives (annotated with SNMIFRE, CIM-10, MSHFRE: T061 T184 T061 T023 or T061 T184 T061 T191
Is (T061 T184 T061) "more semantically similar" to T023 or than T191 ?
Any ideas ?