You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
One of the crucial topics as I see it - is to create a "Standard template" so that we could compare against it
existing standards;
enrichment results;
new standards.
The idea is that it will help to fill all necessary fields and ensure that all objects following same predicates have the same datatype (URI / String / Date / ...).
I made an attempt to create such a template in JSON format and as an image:
On the image above, red area consists of predicates which are necessary for each standard. Yellow - useful predicates that can help to find out more about it and to enrich the ontology if needed. Green - predicates, providing additional information. Inside parenthesis you can see the datatypes of each predicate.
Comparison against the template was implemented here. While checking the structure, this piece of code saves its logs to the file. Here you can find automatically generated logs after applying the approach to the original ontology.
From my point of view, the hierarchy of predicates along with their datatypes should be carefully revised.
The text was updated successfully, but these errors were encountered:
This is definitely a great idea. First of all, the template SHOULD NOT be in a JSON file. This checking should be made based on the STO ontology. We should ask for all the properties in which the Standard class is the domain. Second, all the predicates should be carefully checked especially the mandatory predicates. Some properties are missing: sto:officialResource, sto:hasClassification.
JSON is the simplest implementation of the "hierarchical" (necessary-useful-additional) structure of the Standards' predicates. In addition, it is easy to create a visualisation of the "hierarchy", based on JSON.
Unfortunately, at the current state, STO is not fully consistent. Here you can find some statistics regarding the usage of predicates. Only rdf:type and rdfs:label are used in 100 out of 100 standards. That's why it is hard to rely only on STO.
Regarding missing predicates - you can find sto:hasOfficialResource at the first row in the yellow layer and sto:hasClassification at the second row in the yellow layer on the image above.
One of the crucial topics as I see it - is to create a "Standard template" so that we could compare against it
The idea is that it will help to fill all necessary fields and ensure that all objects following same predicates have the same datatype (URI / String / Date / ...).
I made an attempt to create such a template in JSON format and as an image:
On the image above, red area consists of predicates which are necessary for each standard. Yellow - useful predicates that can help to find out more about it and to enrich the ontology if needed. Green - predicates, providing additional information. Inside parenthesis you can see the datatypes of each predicate.
Comparison against the template was implemented here. While checking the structure, this piece of code saves its logs to the file. Here you can find automatically generated logs after applying the approach to the original ontology.
From my point of view, the hierarchy of predicates along with their datatypes should be carefully revised.
The text was updated successfully, but these errors were encountered: