Build gold standard for evaluation
marfox opened this issue · comments
A fully annotated set of 500 sentences is needed to evaluate the classifiers.
We still have to double-check some 60 sentences.
dbpedia / fact-extractor
Fact Extraction from Wikipedia Text