Skip to main content

Table 7 Coverage achieved when the estimated coverage reached 99% (assuming the named entities of the other categories are already annotated in the corpus)

From: Accelerating the annotation of sparse named entities by dynamic sentence selection

  Coverage # Sentences Annotated Percentage in the Corpus
CoNLL: LOC 98.5% 5,500 39.2%
CoNLL: MISC 95.0% 3,200 22.8%
CoNLL: ORG 99.0% 5,400 38.5%
CoNLL: PER 97.9% 4,700 33.5%
GENIA: DNA 99.6% 8,200 44.2%
GENIA: RNA 99.5% 1,800 9.7%
GENIA: cell_line 99.3% 5,000 27.0%
GENIA: cell_type 99.2% 7,000 37.7%
Average 98.5% - 31.6%