home edit page issue tracker

This page pertains to UD version 2.

Treebank Statistics: UD_Cantonese-HK: POS Tags: ADJ

There are 71 ADJ lemmas (7%), 143 ADJ types (8%) and 373 ADJ tokens (3%). Out of 15 observed tags, the rank of ADJ is: 4 in number of lemmas, 4 in number of types and 8 in number of tokens.

The 10 most frequent ADJ lemmas: _、 好、 快、 多、 高、 得閒、 啱、 大、 新、 第二

The 10 most frequent ADJ types: 好、 第一、 第二、 多、 低、 大、 快、 啱、 高、 得閒

The 10 most frequent ambiguous lemmas: _ (PUNCT 1377, VERB 1352, NOUN 1283, ADV 853, PART 764, PRON 662, AUX 335, DET 217, ADJ 209, ADP 140, NUM 124, SCONJ 101, CCONJ 93, INTJ 92, PROPN 52), 好 (ADV 34, ADJ 22, AUX 1), 快 (ADJ 12, ADV 9), 多 (ADJ 11, DET 1), 新 (ADJ 5, ADV 1), 細 (ADJ 3, VERB 1), mean (ADJ 1, NOUN 1), 夜 (ADJ 1, NOUN 1), 大聲 (ADV 2, ADJ 1), 興 (ADJ 1, VERB 1)

The 10 most frequent ambiguous types: 好 (ADV 54, ADJ 37, AUX 4), 多 (ADJ 14, DET 3), 快 (ADJ 12, ADV 9), 啱 (ADJ 11, VERB 1), 新 (ADJ 6, ADV 1), 正式 (ADJ 4, ADV 3), 錯 (ADJ 4, NOUN 3), 明 (ADJ 3, PROPN 2), 有關 (ADP 7, ADJ 3), 正 (ADJ 3, VERB 2)

Morphology

The form / lemma ratio of ADJ is 2.014085 (the average of all parts of speech is 1.624294).

The 1st highest number of forms (86) was observed with the lemma “_”: 一樣, 不如, 中立, 代理, 低, 個人, 充足, 先後, 公平, 公正, 公眾, 公道, 公開, 具體, 原來, 原裝, 口頭上面, 合法, 合理, 啱, 嚴謹, 嚴重, 困難, 在坐, 多, 夠, 大, 好, 完整, 完絕, 容易, 對, 成功, 新, 既定, 明, 書面, 最後, 有效, 有關, 正式, 正正式式, 正確, 正規, 歷屆, 混亂, 清, 清晰, 清楚, 清醒, 滿意, 特別, 直選, 相關, 真, 真真正正, 稍後, 第一, 第七, 第三, 第二, 第六, 第十二, 第四, 第四十二, 簡單, 粗疏, 糟糕, 緊, 緊要, 英籍, 草率, 莊嚴, 謹慎, 足夠, 適宜, 重大, 重要, 錯, 難, 非宗教, 非建制派, 類似, 騷擾, 高, 點樣.

The 2nd highest number of forms (2) was observed with the lemma “重”: 仲, 重.

The 3rd highest number of forms (1) was observed with the lemma “Bad”: Bad.

ADJ does not occur with any features.

Relations

ADJ nodes are attached to their parents using 18 different relations: amod (122; 33% instances), root (74; 20% instances), conj (39; 10% instances), advmod (30; 8% instances), compound:vv (29; 8% instances), ccomp (15; 4% instances), xcomp (15; 4% instances), advcl (11; 3% instances), parataxis (10; 3% instances), acl (7; 2% instances), obj (5; 1% instances), csubj (4; 1% instances), nmod (3; 1% instances), reparandum (3; 1% instances), compound (2; 1% instances), obl:tmod (2; 1% instances), nsubj (1; 0% instances), obl (1; 0% instances)

Parents of ADJ nodes belong to 10 different parts of speech: NOUN (130; 35% instances), VERB (126; 34% instances), (74; 20% instances), ADJ (33; 9% instances), ADV (3; 1% instances), PRON (2; 1% instances), PROPN (2; 1% instances), ADP (1; 0% instances), AUX (1; 0% instances), DET (1; 0% instances)

128 (34%) ADJ nodes are leaves.

95 (25%) ADJ nodes have one child.

56 (15%) ADJ nodes have two children.

94 (25%) ADJ nodes have three or more children.

The highest child degree of a ADJ node is 12.

Children of ADJ nodes are attached using 33 different relations: advmod (185; 28% instances), punct (131; 20% instances), discourse:sp (58; 9% instances), conj (45; 7% instances), nsubj (45; 7% instances), mark:rel (36; 6% instances), cop (17; 3% instances), parataxis (17; 3% instances), advcl (12; 2% instances), aux (12; 2% instances), discourse (12; 2% instances), case (10; 2% instances), clf (9; 1% instances), reparandum (7; 1% instances), obl:tmod (6; 1% instances), cc (5; 1% instances), csubj (5; 1% instances), mark:adv (5; 1% instances), advcl:coverb (4; 1% instances), mark (4; 1% instances), obl (4; 1% instances), vocative (4; 1% instances), nsubj:periph (3; 0% instances), xcomp (3; 0% instances), compound:ext (2; 0% instances), compound:vv (2; 0% instances), nummod (2; 0% instances), ccomp (1; 0% instances), compound (1; 0% instances), compound:quant (1; 0% instances), det (1; 0% instances), dislocated (1; 0% instances), nmod (1; 0% instances)

Children of ADJ nodes belong to 15 different parts of speech: ADV (195; 30% instances), PUNCT (131; 20% instances), PART (102; 16% instances), NOUN (57; 9% instances), VERB (54; 8% instances), ADJ (33; 5% instances), AUX (28; 4% instances), PRON (19; 3% instances), INTJ (9; 1% instances), CCONJ (7; 1% instances), ADP (5; 1% instances), PROPN (4; 1% instances), SCONJ (3; 0% instances), DET (2; 0% instances), NUM (2; 0% instances)