- From: Piero Bonatti <pieroandrea.bonatti@unina.it>
- Date: Fri, 13 Dec 2019 16:29:28 +0100
- To: public-dpvcg@w3.org
Hallo Harsh and Xiaohu, concerning historical data, may I support the view expressed in the mentioned chat, that is: On 13/12/19 13:42, Harshvardhan J. Pandit wrote: > If one wants to express health data [or any other data category] > as also historial data, they can > subclass it > > e.g. HealthHistory rdfs:subClassOf HealthData, HistoricalData. The same approach would apply, by analogy, to "AnonymizedData" (a useful term that IMHO is currently missing in the vocabulary). Note that "anonymized" is not the same as "anonymous", since all the known anonymization methods give only probabilistic guarantees of non-identification, therefore they do not imply that the result of anonymization is not "personal data" in the sense of the GDPR. Anonymized data should be regarded as somewhere in between pseudonymous and anonymous. OWL2 and most of its profiles (including the logic PL adopted by special) support also an alternative formulation for expressing the intersection of two classes on the fly, without adding terms to the ontology, for example: ObjectIntersectionOf( HealthData HistoricalData ) ObjectIntersectionOf( HealthData AnonymizedData ) In the aforementioned languages, vocabularies need not be extended with all the intersection of the above form, preventing class proliferation in the vocabulary. regards, Piero
Received on Friday, 13 December 2019 15:29:21 UTC