home edit page issue tracker

This page pertains to UD version 2.

Treebank Statistics: UD_English-EWT: POS Tags: NUM

There are 1228 NUM lemmas (7%), 1243 NUM types (5%) and 5051 NUM tokens (2%). Out of 17 observed tags, the rank of NUM is: 5 in number of lemmas, 5 in number of types and 13 in number of tokens.

The 10 most frequent NUM lemmas: one, two, 2, 1, 3, 5, 4, 10, million, three

The 10 most frequent NUM types: one, two, 2, 1, 3, 5, 4, 10, three, 20

The 10 most frequent ambiguous lemmas: one (NUM 426, NOUN 149, PRON 49), 3 (NUM 140, NOUN 1), million (NUM 83, NOUN 8), 12 (NUM 38, NOUN 1), billion (NUM 35, NOUN 5), 13 (NUM 27, NOUN 1), 16 (NUM 27, NOUN 3), thousand (NUM 24, NOUN 20), 22 (NUM 20, NOUN 1), ten (NUM 17, NOUN 8)

The 10 most frequent ambiguous types: one (NUM 371, NOUN 105, PRON 46, X 3), 2 (NUM 178, ADP 1, PART 1), 4 (NUM 109, ADP 1, SCONJ 1), 20 (NUM 66, X 3), m (NUM 41, AUX 23, NOUN 11, PROPN 3), million (NUM 37, NOUN 1), five (NUM 24, PROPN 1), k (NUM 14, NOUN 3, PROPN 1), b (NUM 16, NOUN 5, AUX 1), ten (NUM 10, NOUN 1)

Morphology

The form / lemma ratio of NUM is 1.012215 (the average of all parts of speech is 1.236902).

The 1st highest number of forms (3) was observed with the lemma “billion”: b, billion, bn.

The 2nd highest number of forms (2) was observed with the lemma “’72”: ‘72, ’72.

The 3rd highest number of forms (2) was observed with the lemma “’73”: ‘73, ’73.

NUM occurs with 4 features: NumForm (5036; 100% instances), NumType (5036; 100% instances), Abbr (84; 2% instances), Typo (1; 0% instances)

NUM occurs with 7 feature-value pairs: Abbr=Yes, NumForm=Digit, NumForm=Roman, NumForm=Word, NumType=Card, NumType=Frac, Typo=Yes

NUM occurs with 7 feature combinations. The most frequent feature combination is NumForm=Digit|NumType=Card (3777 tokens). Examples: 2, 1, 3, 5, 4, 10, 20, 6, 2005, 2003

Relations

NUM nodes are attached to their parents using 30 different relations: nummod (3097; 61% instances), root (425; 8% instances), nmod (291; 6% instances), obl (244; 5% instances), compound (225; 4% instances), appos (179; 4% instances), list (147; 3% instances), obj (100; 2% instances), nsubj (93; 2% instances), conj (90; 2% instances), nmod:tmod (58; 1% instances), parataxis (17; 0% instances), amod (11; 0% instances), xcomp (10; 0% instances), obl:tmod (9; 0% instances), advcl (7; 0% instances), ccomp (7; 0% instances), flat (7; 0% instances), obl:npmod (7; 0% instances), nmod:npmod (6; 0% instances), nsubj:outer (5; 0% instances), nsubj:pass (5; 0% instances), obl:agent (3; 0% instances), acl:relcl (2; 0% instances), dep (1; 0% instances), iobj (1; 0% instances), nmod:poss (1; 0% instances), orphan (1; 0% instances), reparandum (1; 0% instances), vocative (1; 0% instances)

Parents of NUM nodes belong to 11 different parts of speech: NOUN (2345; 46% instances), PROPN (877; 17% instances), VERB (482; 10% instances), NUM (451; 9% instances), (425; 8% instances), SYM (376; 7% instances), ADJ (56; 1% instances), X (17; 0% instances), ADV (14; 0% instances), PRON (5; 0% instances), DET (3; 0% instances)

2795 (55%) NUM nodes are leaves.

1548 (31%) NUM nodes have one child.

513 (10%) NUM nodes have two children.

195 (4%) NUM nodes have three or more children.

The highest child degree of a NUM node is 8.

Children of NUM nodes are attached using 31 different relations: punct (983; 30% instances), case (557; 17% instances), nmod (344; 10% instances), advmod (212; 6% instances), appos (212; 6% instances), nmod:tmod (202; 6% instances), compound (176; 5% instances), conj (101; 3% instances), nummod (90; 3% instances), cop (88; 3% instances), nsubj (84; 3% instances), cc (82; 2% instances), parataxis (46; 1% instances), det (43; 1% instances), amod (24; 1% instances), acl:relcl (14; 0% instances), mark (11; 0% instances), obl (11; 0% instances), aux (10; 0% instances), nmod:npmod (9; 0% instances), acl (6; 0% instances), discourse (5; 0% instances), advcl (4; 0% instances), det:predet (2; 0% instances), nmod:poss (2; 0% instances), reparandum (2; 0% instances), cc:preconj (1; 0% instances), csubj (1; 0% instances), flat (1; 0% instances), orphan (1; 0% instances), vocative (1; 0% instances)

Children of NUM nodes belong to 17 different parts of speech: PUNCT (983; 30% instances), NOUN (599; 18% instances), ADP (467; 14% instances), NUM (451; 14% instances), ADV (174; 5% instances), SYM (120; 4% instances), ADJ (100; 3% instances), AUX (99; 3% instances), CCONJ (79; 2% instances), PRON (76; 2% instances), VERB (61; 2% instances), DET (52; 2% instances), PROPN (45; 1% instances), SCONJ (8; 0% instances), PART (7; 0% instances), INTJ (3; 0% instances), X (1; 0% instances)