1

I'm looking for an ontology that describes the theme of a text (eg. "finance", "leisure"). A text can have multiple themes. Preferably the ontology would also have room to store a match percentage:

{
    "theme": "finance",
    "match": "96.2%"
}

I looked at foaf:theme but it appears to be deprecated and doesn't allow any extra properties.

1
  • For a straight theme, use <doc> dc:subject "foo"
  • To add confidence, you can use RDF*:
    • << <doc> dc:subject "foo" >> my:confidence 0.962

But maybe RDF* is an overkill.

  • You can use the NIF (NLP Interchange Format) ontology. Use its:taIdentRef to point to the theme (has to be a URL), and I think its:taIdentConf is for the confidence
  • Another option is to use the Web Annotation ontology. Put the doc URL in oa:hasTarget, the theme in oa:hasBody and the confidence in . This is how Ontotext's TAG works (disclaimer: I work for them)

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.