anuzzolese / oke-challenge

19 stars 8 forks source link

Maybe mention extraction errors in task-1 training set #20

Closed jplu closed 9 years ago

jplu commented 9 years ago

Hi,

For this one I'm not sure if it is an error or not. For this mention:

<http://www.ontologydesignpatterns.org/data/oke-challenge/task-1/sentence-86#char=23,31>
        a                     nif:String , nif:RFC5147String ;
        nif:anchorOf          "American"@en ;
        nif:beginIndex        "23"^^xsd:int ;
        nif:endIndex          "31"^^xsd:int ;
        nif:referenceContext  <http://www.ontologydesignpatterns.org/data/oke-challenge/task-1/sentence-86#char=0,220> ;
        itsrdf:taIdentRef     oke:American_poet .

The sentence is:

Cleopatra Mathis is an American poet who since 1982 has been the Frederick Sessions Beebe Professor in the English department at Dartmouth College, where Cleopatra Mathis is also director of the Creative Writing Program.

Only "American" is extracted, "poet" should be extracted also as a separate mention, no ? But it depends of the decision you took for #8.

If my previous guess is correct why doing these two extractions:

<http://www.ontologydesignpatterns.org/data/oke-challenge/task-1/sentence-42#char=26,34>
        a                     nif:String , nif:RFC5147String ;
        nif:anchorOf          "American"@en ;
        nif:beginIndex        "26"^^xsd:int ;
        nif:endIndex          "34"^^xsd:int ;
        nif:referenceContext  <http://www.ontologydesignpatterns.org/data/oke-challenge/task-1/sentence-42#char=0,76> ;
        itsrdf:taIdentRef     oke:American .

And

<http://www.ontologydesignpatterns.org/data/oke-challenge/task-1/sentence-42#char=35,54>
        a                     nif:String , nif:RFC5147String ;
        nif:anchorOf          "Methodist Episcopal"@en ;
        nif:beginIndex        "35"^^xsd:int ;
        nif:endIndex          "54"^^xsd:int ;
        nif:referenceContext  <http://www.ontologydesignpatterns.org/data/oke-challenge/task-1/sentence-42#char=0,76> ;
        itsrdf:taIdentRef     oke:Methodist_Episcopal .

In the sentence:

Henry Herman Meyer was an American Methodist Episcopal clergyman and editor.

Instead of:

<http://www.ontologydesignpatterns.org/data/oke-challenge/task-1/sentence-42#char=26,54>
        a                     nif:String , nif:RFC5147String ;
        nif:anchorOf          "American Methodist Episcopal"@en ;
        nif:beginIndex        "26"^^xsd:int ;
        nif:endIndex          "54"^^xsd:int ;
        nif:referenceContext  <http://www.ontologydesignpatterns.org/data/oke-challenge/task-1/sentence-42#char=0,76> ;
        itsrdf:taIdentRef     oke:American_Methodist_Episcopal .

Mainly because you have annotated those two differents mentions as a "Person".

Cheers.

jplu commented 9 years ago

I close this issue as it as to be divided in two different one.