Treebank Statistics: UD_Turkish-Penn: POS Tags: PROPN
There are 5019 PROPN lemmas (32%), 6343 PROPN types (17%) and 15486 PROPN tokens (8%).
Out of 15 observed tags, the rank of PROPN is: 2 in number of lemmas, 3 in number of types and 4 in number of tokens.
The 10 most frequent PROPN lemmas: abd, new, york, corp., co., dow, ınc., japonya, jones, wall
The 10 most frequent PROPN types: New, ABD, York, Corp., Dow, Jones, Wall, Co., Inc., Inc
The 10 most frequent ambiguous lemmas: abd (PROPN 259, NOUN 6), cuma (NOUN 79, PROPN 64), ekim (PROPN 60, NOUN 43), federal (ADJ 83, PROPN 54), kasım (PROPN 54, NOUN 24), eylül (PROPN 49, NOUN 44), ağustos (PROPN 45, NOUN 31), yen (NOUN 62, PROPN 45, VERB 3), doğu (PROPN 42, ADJ 10, NOUN 5), san (PROPN 41, VERB 34, NOUN 7)
The 10 most frequent ambiguous types: ABD (PROPN 189, NOUN 6), Cuma (PROPN 60, NOUN 47), Federal (PROPN 48, ADJ 25), San (PROPN 41, NOUN 7), Ekim (PROPN 40, NOUN 34), Eylül (PROPN 40, NOUN 34), Beyaz (PROPN 35, ADJ 5), Güney (PROPN 32, NOUN 11, ADJ 10), Kıymetler (NOUN 56, PROPN 32), Doğu (PROPN 30, ADJ 8, NOUN 1)
- ABD
- Cuma
- Federal
- San
- Ekim
- Eylül
- Beyaz
- Güney
- Kıymetler
- Doğu
Morphology
The form / lemma ratio of PROPN is 1.263798 (the average of all parts of speech is 2.343544).
The 1st highest number of forms (9) was observed with the lemma “abd”: ABD, ABD’de, ABD’deki, ABD’den, ABD’dir, ABD’li, ABD’nin, ABD’ye, ABD’yi.
The 2nd highest number of forms (8) was observed with the lemma “co.”: Co., Co.’da, Co.’daki, Co.’dan, Co.’dur, Co.’nun, Co.’su, Co.’yu.
The 3rd highest number of forms (8) was observed with the lemma “francisco”: Francisco, Francisco’da, Francisco’daki, Francisco’dan, Francisco’lu, Francisco’ludan, Francisco’nun, Francisco’ya.
PROPN occurs with 4 features: Number (15218; 98% instances), Case (15185; 98% instances), Number[psor] (87; 1% instances), Typo (14; 0% instances)
PROPN occurs with 12 feature-value pairs: Case=Abl, Case=Acc, Case=Dat, Case=Gen, Case=Ins, Case=Loc, Case=Nom, Number=Plur, Number=Sing, Number[psor]=Plur, Number[psor]=Sing, Typo=Yes
PROPN occurs with 27 feature combinations.
The most frequent feature combination is Case=Nom|Number=Sing (12588 tokens).
Examples: New, ABD, York, Corp., Dow, Jones, Wall, Inc., Co., Inc
Relations
PROPN nodes are attached to their parents using 22 different relations: nmod (4951; 32% instances), flat (3757; 24% instances), nsubj (3310; 21% instances), obl (983; 6% instances), conj (485; 3% instances), compound (442; 3% instances), amod (431; 3% instances), root (397; 3% instances), obj (285; 2% instances), appos (258; 2% instances), list (125; 1% instances), advcl (16; 0% instances), parataxis (10; 0% instances), xcomp (10; 0% instances), vocative (7; 0% instances), ccomp (4; 0% instances), discourse (4; 0% instances), acl (3; 0% instances), iobj (3; 0% instances), clf (2; 0% instances), csubj (2; 0% instances), nsubj:outer (1; 0% instances)
Parents of PROPN nodes belong to 14 different parts of speech: PROPN (5122; 33% instances), NOUN (5078; 33% instances), VERB (4188; 27% instances), (397; 3% instances), ADJ (363; 2% instances), NUM (137; 1% instances), ADV (117; 1% instances), X (31; 0% instances), PRON (23; 0% instances), INTJ (14; 0% instances), CCONJ (7; 0% instances), ADP (5; 0% instances), DET (3; 0% instances), AUX (1; 0% instances)
7723 (50%) PROPN nodes are leaves.
5965 (39%) PROPN nodes have one child.
1225 (8%) PROPN nodes have two children.
573 (4%) PROPN nodes have three or more children.
The highest child degree of a PROPN node is 9.
Children of PROPN nodes are attached using 26 different relations: flat (3806; 36% instances), nmod (2066; 20% instances), punct (1248; 12% instances), conj (543; 5% instances), case (439; 4% instances), cc (421; 4% instances), compound (386; 4% instances), amod (363; 3% instances), nummod (287; 3% instances), acl (270; 3% instances), list (153; 1% instances), appos (146; 1% instances), det (82; 1% instances), advmod (81; 1% instances), nsubj (73; 1% instances), obl (29; 0% instances), goeswith (14; 0% instances), advcl (10; 0% instances), discourse (9; 0% instances), mark (9; 0% instances), parataxis (8; 0% instances), aux (7; 0% instances), ccomp (2; 0% instances), clf (2; 0% instances), obj (2; 0% instances), csubj (1; 0% instances)
Children of PROPN nodes belong to 14 different parts of speech: PROPN (5122; 49% instances), NOUN (2092; 20% instances), PUNCT (1248; 12% instances), CCONJ (564; 5% instances), NUM (361; 3% instances), VERB (304; 3% instances), ADJ (246; 2% instances), ADP (229; 2% instances), ADV (140; 1% instances), DET (86; 1% instances), X (25; 0% instances), PRON (18; 0% instances), INTJ (15; 0% instances), AUX (7; 0% instances)