Treebank Statistics: UD_English-GUM: Features: NumForm
This feature is language-specific.
It occurs with 3 different values: Digit, Roman, Word.
3265 tokens (2%) have a non-empty value of NumForm.
622 types (4%) occur at least once with a non-empty value of NumForm.
596 lemmas (5%) occur at least once with a non-empty value of NumForm.
The feature is used with 5 part-of-speech tags: NUM (3175; 2% instances), NOUN (44; 0% instances), ADV (37; 0% instances), DET (7; 0% instances), PROPN (2; 0% instances).
NUM
3175 NUM tokens (100% of all NUM tokens) have a non-empty value of NumForm.
The most frequent other feature values with which NUM and NumForm co-occurred: NumType=Card (3094; 97%).
NUM tokens may have the following values of NumForm:
Digit(2193; 69% of non-emptyNumForm): 1, 2, 3, 6, 4, 10, 5, 15, 7, 12Roman(17; 1% of non-emptyNumForm): II, I, III, XV, XVIIWord(965; 30% of non-emptyNumForm): one, two, three, four, five, six, million, ten, eight, twenty
NumForm seems to be lexical feature of NUM. 100% lemmas (588) occur only with one value of NumForm.
NOUN
44 NOUN tokens (0% of all NOUN tokens) have a non-empty value of NumForm.
The most frequent other feature values with which NOUN and NumForm co-occurred: Number=Sing (42; 95%).
NOUN tokens may have the following values of NumForm:
Word(44; 100% of non-emptyNumForm): half, quarter, third, millionth, quarters, tenth, thirds
ADV
37 ADV tokens (0% of all ADV tokens) have a non-empty value of NumForm.
The most frequent other feature values with which ADV and NumForm co-occurred: PronType=EMPTY (37; 100%), Degree=EMPTY (35; 95%).
ADV tokens may have the following values of NumForm:
Word(37; 100% of non-emptyNumForm): once, twice, half
DET
7 DET tokens (0% of all DET tokens) have a non-empty value of NumForm.
The most frequent other feature values with which DET and NumForm co-occurred: Definite=EMPTY (7; 100%), PronType=Ind (7; 100%).
DET tokens may have the following values of NumForm:
Word(7; 100% of non-emptyNumForm): half
PROPN
2 PROPN tokens (0% of all PROPN tokens) have a non-empty value of NumForm.
The most frequent other feature values with which PROPN and NumForm co-occurred: Number=Sing (2; 100%).
PROPN tokens may have the following values of NumForm:
Word(2; 100% of non-emptyNumForm): EIGHT, One
Relations with Agreement in NumForm
The 10 most frequent relations where parent and child node agree in NumForm:
NUM –[nmod:tmod]–> NUM (122; 100%),
NUM –[nmod]–> NUM (108; 98%),
NUM –[conj]–> NUM (100; 94%),
NUM –[compound]–> NUM (40; 65%),
NUM –[flat]–> NUM (11; 100%),
NUM –[parataxis]–> NUM (7; 100%),
NUM –[dep]–> NUM (5; 83%),
NUM –[appos]–> NUM (1; 100%),
NUM –[det:predet]–> DET (1; 100%),
NUM –[reparandum]–> NUM (1; 100%).