home edit page issue tracker

This page pertains to UD version 2.

Treebank Statistics: UD_Estonian-EDT: POS Tags: NUM

There are 1338 NUM lemmas (4%), 1581 NUM types (2%) and 7407 NUM tokens (2%). Out of 16 observed tags, the rank of NUM is: 6 in number of lemmas, 6 in number of types and 11 in number of tokens.

The 10 most frequent NUM lemmas: kaks, üks, kolm, miljon, 1, viis, neli, 2, 000, kümme

The 10 most frequent NUM types: kaks, 1, 000, 2, üks, kolm, kahe, ühe, 3, miljonit

The 10 most frequent ambiguous lemmas: üks (DET 641, NUM 344, PRON 112), viis (NUM 155, NOUN 50), pool (NUM 128, NOUN 89, ADV 28, ADP 13), paar (NUM 124, NOUN 17), 2000 (NUM 79, ADJ 1), seitse (NUM 65, NOUN 1), sada (NUM 60, NOUN 2), 2001 (NUM 53, ADJ 1), 17 (NUM 44, ADJ 1), kolmandik (NUM 21, NOUN 7)

The 10 most frequent ambiguous types: üks (DET 180, NUM 118, PRON 41), ühe (DET 101, NUM 99, PRON 7), viis (NUM 77, VERB 34, NOUN 5), 2000 (NUM 78, ADJ 1), paar (NUM 58, NOUN 5), 2001 (NUM 53, ADJ 1), pool (NUM 44, ADV 28, NOUN 14, ADP 12), poole (ADP 139, NUM 46, NOUN 15, ADV 3), 17 (NUM 44, ADJ 1), seitse (NUM 36, NOUN 1)

Morphology

The form / lemma ratio of NUM is 1.181614 (the average of all parts of speech is 1.912184).

The 1st highest number of forms (19) was observed with the lemma “üks”: ühe, ühe-, ühega, ühegi, üheks, ühel, ühele, ühelgi, ühelt, ühena, üheni, ühes, ühest, üht, ühte, ühtegi, ühtki, üks, ükski.

The 2nd highest number of forms (14) was observed with the lemma “kümme”: kümme, kümmet, kümne, kümned, kümnega, kümneid, kümneks, kümnel, kümnele, kümnest, kümnete, kümnetes, kümnetesse, kümnetest.

The 3rd highest number of forms (13) was observed with the lemma “miljon”: Miljonitel, miljon, miljoneid, miljoni, miljonid, miljoniga, miljonilt, miljonini, miljonist, miljonit, miljonite, miljoniteni, miljonitest.

NUM occurs with 7 features: NumType (7407; 100% instances), NumForm (7264; 98% instances), Case (5696; 77% instances), Number (5696; 77% instances), Hyph (42; 1% instances), Abbr (5; 0% instances), PronType (2; 0% instances)

NUM occurs with 24 feature-value pairs: Abbr=Yes, Case=Abl, Case=Add, Case=Ade, Case=All, Case=Com, Case=Ela, Case=Ess, Case=Gen, Case=Ill, Case=Ine, Case=Nom, Case=Par, Case=Ter, Case=Tra, Hyph=Yes, NumForm=Digit, NumForm=Letter, NumForm=Roman, NumType=Card, NumType=Ord, Number=Plur, Number=Sing, PronType=Ind

NUM occurs with 60 feature combinations. The most frequent feature combination is Case=Nom|Number=Sing|NumForm=Digit|NumType=Card (2244 tokens). Examples: 1, 000, 2, 2000, 3, 4, 15, 10, 5, 1999

Relations

NUM nodes are attached to their parents using 22 different relations: nummod (5225; 71% instances), parataxis (615; 8% instances), compound (456; 6% instances), conj (391; 5% instances), root (157; 2% instances), flat (136; 2% instances), nsubj (122; 2% instances), obj (74; 1% instances), obl (55; 1% instances), nsubj:cop (54; 1% instances), goeswith (34; 0% instances), appos (31; 0% instances), acl:relcl (13; 0% instances), orphan (9; 0% instances), amod (8; 0% instances), nmod (8; 0% instances), advcl (7; 0% instances), advmod (4; 0% instances), ccomp (3; 0% instances), csubj:cop (3; 0% instances), csubj (1; 0% instances), list (1; 0% instances)

Parents of NUM nodes belong to 11 different parts of speech: NOUN (4556; 62% instances), NUM (980; 13% instances), VERB (884; 12% instances), PROPN (517; 7% instances), ADJ (190; 3% instances), (157; 2% instances), ADV (91; 1% instances), SYM (18; 0% instances), PRON (9; 0% instances), X (4; 0% instances), CCONJ (1; 0% instances)

4407 (59%) NUM nodes are leaves.

1844 (25%) NUM nodes have one child.

801 (11%) NUM nodes have two children.

355 (5%) NUM nodes have three or more children.

The highest child degree of a NUM node is 11.

Children of NUM nodes are attached using 27 different relations: punct (1843; 37% instances), advmod (586; 12% instances), compound (445; 9% instances), nmod (439; 9% instances), conj (411; 8% instances), cc (167; 3% instances), case (158; 3% instances), amod (135; 3% instances), nsubj:cop (120; 2% instances), cop (116; 2% instances), nummod (112; 2% instances), det (91; 2% instances), obl (87; 2% instances), flat (83; 2% instances), goeswith (37; 1% instances), parataxis (29; 1% instances), mark (19; 0% instances), aux (18; 0% instances), advcl (15; 0% instances), acl:relcl (11; 0% instances), appos (3; 0% instances), cc:preconj (2; 0% instances), orphan (2; 0% instances), xcomp (2; 0% instances), fixed (1; 0% instances), list (1; 0% instances), nsubj (1; 0% instances)

Children of NUM nodes belong to 15 different parts of speech: PUNCT (1843; 37% instances), NUM (980; 20% instances), ADV (619; 13% instances), NOUN (544; 11% instances), ADJ (170; 3% instances), CCONJ (167; 3% instances), ADP (159; 3% instances), AUX (134; 3% instances), DET (92; 2% instances), PROPN (81; 2% instances), PRON (75; 2% instances), VERB (37; 1% instances), SCONJ (17; 0% instances), SYM (13; 0% instances), X (3; 0% instances)