home edit page issue tracker

This page pertains to UD version 2.

Treebank Statistics: UD_Turkish-GB: POS Tags: NUM

There are 55 NUM lemmas (2%), 91 NUM types (2%) and 372 NUM tokens (2%). Out of 16 observed tags, the rank of NUM is: 6 in number of lemmas, 8 in number of types and 11 in number of tokens.

The 10 most frequent NUM lemmas: iki, üç, bir, dört, beş, yüz, on, yedi, kaç, bin

The 10 most frequent NUM types: iki, üç, beş, yüz, dört, on, bir, bin, birer, dokuz

The 10 most frequent ambiguous lemmas: iki (NUM 66, ADJ 5, DET 2), üç (NUM 30, ADJ 2), bir (DET 351, NUM 29, ADV 4, PRON 2, ADJ 1), dört (NUM 27, ADJ 1), yüz (NUM 23, VERB 6, NOUN 5), kaç (NUM 15, VERB 5), bin (NUM 13, VERB 6), milyon (NUM 3, ADJ 1), yarım (NUM 2, ADJ 1)

The 10 most frequent ambiguous types: iki (NUM 48, DET 1), bir (DET 322, NUM 14, ADV 3, X 1), kaç (NUM 9, VERB 1), yedi (NUM 7, VERB 2), İki (NUM 7, ADJ 1, DET 1), biri (PRON 9, NUM 3), milyon (NUM 2, ADJ 1), ikinci (ADJ 3, NUM 1)

Morphology

The form / lemma ratio of NUM is 1.654545 (the average of all parts of speech is 2.332157).

The 1st highest number of forms (7) was observed with the lemma “iki”: iki, ikide, ikiden, ikinci, ikiye, ikişer, İki.

The 2nd highest number of forms (6) was observed with the lemma “yedi”: yedi, yedide, yedisi, yedisinde, yediye, yediyi.

The 3rd highest number of forms (5) was observed with the lemma “bir”: bir, bire, birer, biri, birinde.

NUM occurs with 7 features: NumType (372; 100% instances), Number (78; 21% instances), Case (72; 19% instances), Number[psor] (12; 3% instances), Person[psor] (12; 3% instances), PronType (8; 2% instances), Definite (1; 0% instances)

NUM occurs with 15 feature-value pairs: Case=Abl, Case=Acc, Case=Dat, Case=Gen, Case=Ins, Case=Loc, Case=Nom, Definite=Ind, NumType=Card, NumType=Dist, Number=Plur, Number=Sing, Number[psor]=Sing, Person[psor]=3, PronType=Int

NUM occurs with 18 feature combinations. The most frequent feature combination is NumType=Card (260 tokens). Examples: iki, üç, beş, yüz, dört, on, bir, bin, dokuz, sekiz

Relations

NUM nodes are attached to their parents using 13 different relations: nummod (193; 52% instances), compound (69; 19% instances), root (34; 9% instances), obl:tmod (33; 9% instances), nmod (15; 4% instances), conj (14; 4% instances), nsubj (4; 1% instances), compound:redup (3; 1% instances), obj (2; 1% instances), obl (2; 1% instances), fixed (1; 0% instances), nmod:part (1; 0% instances), nsubj:cop (1; 0% instances)

Parents of NUM nodes belong to 7 different parts of speech: NOUN (206; 55% instances), NUM (93; 25% instances), (34; 9% instances), VERB (33; 9% instances), ADJ (4; 1% instances), ADV (1; 0% instances), PRON (1; 0% instances)

274 (74%) NUM nodes are leaves.

58 (16%) NUM nodes have one child.

22 (6%) NUM nodes have two children.

18 (5%) NUM nodes have three or more children.

The highest child degree of a NUM node is 5.

Children of NUM nodes are attached using 13 different relations: compound (69; 41% instances), nmod (32; 19% instances), case (20; 12% instances), conj (15; 9% instances), nmod:part (7; 4% instances), nummod (6; 4% instances), advmod (5; 3% instances), punct (5; 3% instances), compound:redup (3; 2% instances), nsubj (3; 2% instances), aux (1; 1% instances), cop (1; 1% instances), csubj:cop (1; 1% instances)

Children of NUM nodes belong to 9 different parts of speech: NUM (93; 55% instances), NOUN (39; 23% instances), ADP (20; 12% instances), ADV (5; 3% instances), PUNCT (5; 3% instances), AUX (2; 1% instances), VERB (2; 1% instances), PROPN (1; 1% instances), SYM (1; 1% instances)