Output in triplet form

Hi Team,

I am training my unstructured text file in this form with annotations:

START:brand Monster High are nothing to do with high.
START:brand Exerpeutic is a brand.
START:brand U-knit
START:brand Pepsi
START:brand Platinum Pets are not talking about pets.
Spiderman is like a START:brand Batman .
START:brand ProGear sports outfit is cool stuff.
START:brand Spiderman
START:brand Exerpeutic
START:brand Dickies is also a brand.
START:brand BuySeasons
START:brand Microsoft
START:brand Microsoft is own by Bill Gates.
START:brand Transformers is a movie.
....................
....................

Now I, getting the output in the following format in stardog

I want to see the output in triplet format, like:

ProGear is brand
(we have annotated it as a brand in training file) and testing it on the same file without annotations.

Please suggest how it can be done.

Thanks
Navratan

Hi Navratan,

The type of the entity is saved, as an IRI, with the rdf:type property. In your example, tag:stardog:api:docs:ner:brand.

You can use Stardog to extract the local name of the IRI and get it as a string. e.g.,:

select ?label (localname(?type) as ?typeString) where {
     graph ?g {
        ?doc stardog:docs:hasEntity ?entity .

        ?entity a ?type ;
                rdfs:label ?label .
    }
}

-pedro

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.