Skip to main content
. 2007 Sep-Oct;14(5):641–650. doi: 10.1197/jamia.M2392

Table 6.

Table 6 Comparison of Errors Assigning POS for Tagger Trained with Three Different Corpora

Reference tag → Error ME tagger trained with WSJ only ME tagger trained with WSJ + DSC ME tagger trained with WSJ + H1 selected Data
CC → RB 3 3 3
CD → JJ 0 0 2
CD → NN 1 0 2
CD → NNP 21 0 0
IN → NNP 10 0 0
IN → RB 3 2 2
JJ → NN 77 49 56
JJ → NNP 271 4 7
LS → NNP 70 20 30
NN → JJ 23 28 47
NN → NNP 310 3 3
NNP → JJ 2 9 7
NNS → NN 89 13 8
RB → JJ 7 3 2
VBD → VBN 57 20 20
VBN → NNP 12 0 0
VBN → VBD 10 21 16
VBZ → NNS 14 3 3

CC = conjunction, coordinating; RB = adverb; CD = numeral, cardinal; JJ = adjective; NN = noun, singular or mass; NNP = noun, proper, singular; IN = preposition or conjunction, subordinating; LS = list item marker; NNS = noun, plural; VBD = verb, past tense; VBN = verb, past participle; VBG = verb, present participle or gerund.