Fix overlapping markings in MSNBC dataset
Closed this issue · 1 comments
MichaelRoeder commented
The MSNBC dataset seems to have some markings, that are overlapping.
example: Bus16451112.txt
from 699, to 710, URI=http://dbpedia.org/resource/Frank_Blake
from 705, to 710, URI=http://dbpedia.org/resource/Frank_Blake
MichaelRoeder commented
There is already a span merging inside the org.aksw.gerbil.evaluate.impl.SpanMergingEvaluatorDecorator
class that could be extracted / reused.