Difference between revisions of "Dhivehi/Universal Dependencies"

From LING073
Jump to: navigation, search
(→‎Evaluation)
(→‎Evaluation)
 
(10 intermediate revisions by one other user not shown)
Line 5: Line 5:
 
! model !! corpus !! UAS !! LAS !! # of forms !! # of sentences
 
! model !! corpus !! UAS !! LAS !! # of forms !! # of sentences
 
|-
 
|-
−
| div.nomorph.udpipe || div.annotated.ud.conllu || 81.51% || 68.07% || 119 || 25
+
| div.nomorph.udpipe || div.annotated.ud.conllu || 56.55% || 39.29% || 168 || 27
 
|-
 
|-
−
| div.nomorph.udpipe || div.annotated2.ud.conllu || - || - || - || -
+
| div.nomorph.udpipe || div.annotated2.ud.conllu || 54.02% || 34.48% || 87 || 12
 
|-
 
|-
 
| div.withmorph.udpipe || div.annotated.ud.conllu || 56.55% || 39.29% || 168 || 27
 
| div.withmorph.udpipe || div.annotated.ud.conllu || 56.55% || 39.29% || 168 || 27
 
|-
 
|-
−
| div.withmorph.udpipe || div.annotated2.ud.conllu || - || - || -|| -
+
| div.withmorph.udpipe || div.annotated2.ud.conllu || 54.02% || 34.48% || 87 || 12
 
|}
 
|}
  
−
= Common Dependency Relations =
+
{{comment|It seems odd that the results would be the same with and without morph.  Are you sure this is right?  —[[User:Jwashin1|Jwashin1]] ([[User talk:Jwashin1|talk]]) 14:50, 5 June 2019 (EDT)}}
 +
 
 +
= Dependency Relations =
 +
 
 +
== nsubj ==
 +
Subject of the sentence. It can either be the subject of a verb or an adjective. (Adjectives could be the predicate of a sentence.)
 +
 
 +
* '''Examples:'''
 +
<pre>"<އެކުދިން>"
 +
"ކުދިން" n hum pl def dir @nsubj #1->5
 +
"އެ" det dem deg3 @det #2->1
 +
"<ބިސްތައް>"
 +
"ބިސްތައް" n nhum pl def dir @obj #3->5
 +
"<ވަޅުލައިފިއެވެ>"
 +
"އެވެ" mod @discourse #4->5
 +
"ވަޅުލައިފި" v tv act perf p3 @root #5->0
 +
"<.>"
 +
"." sent @punct #6->5</pre>
 +
 
 +
<pre>"<އޭގެ>"
 +
"އޭގެ" prn pers p3 sg pos @nmod:poss #1->2
 +
"<ނޭފަތް>"
 +
"ނޭފަތް" n nhum sg def dir @nsubj #2->5
 +
"<ވަރަށް>"
 +
"ވަރަށް" adv @advmod #3->5
 +
"<ދިގެވެ>"
 +
"އެވެ" mod @discourse #4->5
 +
"ދިގު" adj @root #5->0
 +
"<.>"
 +
"." sent @punct #6->5</pre>
 +
 
 +
== obj ==
 +
Object of the sentence. It's usually used as the object of a verb.
 +
 
 +
* '''Examples:'''
 +
 
 +
<pre>"<އަހަރެންގެ>"
 +
"އަހަރެންގެ" prn pers p1 sg std gen @nsubj #1->4
 +
"<ދެބޭބެއިން>"
 +
"ބޭބެއިން" n hum pl def dir @obj #2->4
 +
"ދެ" num @nummod #3->2
 +
"<ތިބޭ>"
 +
"ތިބޭ" v tv act pres p3 @root #4->0
 +
"<.>"
 +
"." sent @punct #5->4</pre>
 +
 
 +
<pre>"<އަހަރެން>"
 +
"އަހަރެން" prn pers p1 sg std dir @obj #1->3
 +
"<އެކަނި>"
 +
"އެކަނި" adj @amod #2->1
 +
"<ދުކޮހް>"
 +
"ދުކޮހް" v tv act pres p3 imp @root #3->0
 +
"<ލާ>"
 +
"ލާ" mod @discourse #4->3
 +
"<!>"
 +
"!" sent @punct #5->3</pre>
 +
 
 +
== nmod:poss ==
 +
Possessive noun modifier of the sentence. It's usually a genetive case noun modifying another another noun.
 +
 
 +
* '''Examples:'''
 +
 
 +
<pre>"<ތަރައްގީ>"
 +
"ތަރައްގީ" n nhum sg def dir @nsubj #1->4
 +
"<އަކީ>"
 +
"އަކީ" mod @cop #2->4
 +
"<އާއިލާގެ>"
 +
"އާއިލާ" n nhum sg def gen @nmod:poss #3->4
 +
"<އާމްދަނީ>"
 +
"އާމްދަނީ" n nhum sg def dir @root #4->0
 +
"<އިތުރުވުން>"
 +
"އިތުރުވުން" v iv act pres p3 @amod #5->4
 +
"<.>"
 +
"." sent @punct #6->5</pre>
 +
 
 +
<pre>"<މިފަދަ>"
 +
"މިފަދަ" det dem deg1 @det #1->2
 +
"<އިސްލާޙެއް>"
 +
"އިސްލާޙް" n nhum sg ind dir @nsubj #2->4
 +
"<ވާނީ>"
 +
"ވާނީ" v tv act fut pot @cop #3->4
 +
"<ޒަމާނީ>"
 +
"ޒަމާނީ" adj @root #4->0
 +
"<ޑިމޮކްރެސީގެ>"
 +
"ޑިމޮކްރެސީ" n nhum sg def gen @nmod:poss #5->6
 +
"<ރޫޙާއި>"
 +
"ރޫޙާއި" n nhum sg def soc @obj #6->4
 +
"<ޚިލާފު>"
 +
"ޚިލާފު" post @advmod #7->8
 +
"<އިސްލާޙެއް>"
 +
"އިސްލާޙް" n nhum sg ind dir @nmod #8->6
 +
"<.>"
 +
"." sent @punct #9->4</pre>
 +
 
 +
== discourse ==
 +
Discourse Marker of a sentence. It's usually used to mark the modal words in a sentence. For example, it is very common in Dhivehi to have a sentence ending particle that doesn't mean anything. The discourse relation is used to mark that particle.
 +
 
 +
* '''Examples:'''
 +
 
 +
<pre>"<އަލި>"
 +
"އަލި" np @nsubj #1->7
 +
"<ލަނޑުޖަހާނެ>"
 +
"ލަނޑުޖަހާނެ" v tv act fut p3 @advcl #2->7
 +
"<ވަރަށް>"
 +
"ވަރަށް" adv @advmod #3->4
 +
"<ބައިވަރު>"
 +
"ބައިވަރު" det @det #4->5
 +
"<ފުރުޞަތު>"
 +
"ފުރުޞަތު" n nhum sg def dir @obj #5->7
 +
"<ލިބުނެވެ>"
 +
"އެވެ" mod @discourse #6->7
 +
"ލިބުން" v tv act past p3 @root #7->0
 +
"<.>"
 +
"." sent @punct #8->7</pre>
 +
 
 +
<pre>"<މަދްރަސާ>"
 +
"މަދްރަސާ" n nhum sg def dir @nsubj #1->6
 +
"<ގެއަށްވުރެ>"
 +
"ވުރެ" post @advmod #2->3
 +
"ގެ" n nhum sg def dat @obj #3->6
 +
"<މާ>"
 +
"މާ" adv @advmod #4->6
 +
"<ބޮޑެވެ>"
 +
"އެވެ" mod @discourse #5->6
 +
"ބޮޑު" adj @root #6->0
 +
"<.>"
 +
"." sent @punct #7->6</pre>
 +
 
 +
== amod ==
 +
Adjective modifier of the sentence. It's usually used to mark the adjective that modifies a noun.
 +
 
 +
* '''Examples:'''
 +
 
 +
<pre>"<ކުރަން>"
 +
"ކުރަ" v tv act pres p1 @fixed #1->2
 +
"<ޖެހޭ>"
 +
"ޖެހޭ" v tv act pprs @root #2->0
 +
"<ހޭދަ>"
 +
"ހޭދަ" adj @amod #3->4
 +
"<އިތުރުވުން>"
 +
"އިތުރުވުން" n nhum sg def dir @obj #4->2
 +
"<.>"
 +
"." sent @punct #5->2</pre>
 +
 
 +
<pre>"<މާދަމާ>"
 +
"މާދަމާ" adv @advmod #1->5
 +
"<އަލީ>"
 +
"އަލީ" np @nsubj #2->5
 +
"<ފާހަނަ>"
 +
"ފާހަނަ" n nhum sg def dir @obj #3->5
 +
"<ސާފު>"
 +
"ސާފު" adj @amod #4->5
 +
"<ކުރާނެ>"
 +
"ކުރަ" v tv act fut p3 @root #5->0
 +
"<.>"
 +
"." sent @punct #6->5</pre>
 +
 
 +
[[Category:sp19_UD]]
 +
[[Category:Dhivehi]]

Latest revision as of 18:50, 5 June 2019

Evaluation

model corpus UAS LAS # of forms # of sentences
div.nomorph.udpipe div.annotated.ud.conllu 56.55% 39.29% 168 27
div.nomorph.udpipe div.annotated2.ud.conllu 54.02% 34.48% 87 12
div.withmorph.udpipe div.annotated.ud.conllu 56.55% 39.29% 168 27
div.withmorph.udpipe div.annotated2.ud.conllu 54.02% 34.48% 87 12

It seems odd that the results would be the same with and without morph. Are you sure this is right? —Jwashin1 (talk) 14:50, 5 June 2019 (EDT)

Dependency Relations

nsubj

Subject of the sentence. It can either be the subject of a verb or an adjective. (Adjectives could be the predicate of a sentence.)

  • Examples:
"<އެކުދިން>"
	"ކުދިން" n hum pl def dir @nsubj #1->5
		"އެ" det dem deg3 @det #2->1
"<ބިސްތައް>"
	"ބިސްތައް" n nhum pl def dir @obj #3->5
"<ވަޅުލައިފިއެވެ>"
	"އެވެ" mod @discourse #4->5
		"ވަޅުލައިފި" v tv act perf p3 @root #5->0
"<.>"
	"." sent @punct #6->5
"<އޭގެ>"
	"އޭގެ" prn pers p3 sg pos @nmod:poss #1->2
"<ނޭފަތް>"
	"ނޭފަތް" n nhum sg def dir @nsubj #2->5
"<ވަރަށް>"
	"ވަރަށް" adv @advmod #3->5
"<ދިގެވެ>"
	"އެވެ" mod @discourse #4->5
		"ދިގު" adj @root #5->0
"<.>"
	"." sent @punct #6->5

obj

Object of the sentence. It's usually used as the object of a verb.

  • Examples:
"<އަހަރެންގެ>"
	"އަހަރެންގެ" prn pers p1 sg std gen @nsubj #1->4
"<ދެބޭބެއިން>"
	"ބޭބެއިން" n hum pl def dir @obj #2->4
		"ދެ" num @nummod #3->2
"<ތިބޭ>"
	"ތިބޭ" v tv act pres p3 @root #4->0
"<.>"
	"." sent @punct #5->4
"<އަހަރެން>"
	"އަހަރެން" prn pers p1 sg std dir @obj #1->3
"<އެކަނި>"
	"އެކަނި" adj @amod #2->1
"<ދުކޮހް>"
	"ދުކޮހް" v tv act pres p3 imp @root #3->0
"<ލާ>"
	"ލާ" mod @discourse #4->3
"<!>"
	"!" sent @punct #5->3

nmod:poss

Possessive noun modifier of the sentence. It's usually a genetive case noun modifying another another noun.

  • Examples:
"<ތަރައްގީ>"
	"ތަރައްގީ" n nhum sg def dir @nsubj #1->4
"<އަކީ>"
	"އަކީ" mod @cop #2->4
"<އާއިލާގެ>"
	"އާއިލާ" n nhum sg def gen @nmod:poss #3->4
"<އާމްދަނީ>"
	"އާމްދަނީ" n nhum sg def dir @root #4->0
"<އިތުރުވުން>"
	"އިތުރުވުން" v iv act pres p3 @amod #5->4
"<.>"
	"." sent @punct #6->5
"<މިފަދަ>"
	"މިފަދަ" det dem deg1 @det #1->2
"<އިސްލާޙެއް>"
	"އިސްލާޙް" n nhum sg ind dir @nsubj #2->4
"<ވާނީ>"
	"ވާނީ" v tv act fut pot @cop #3->4
"<ޒަމާނީ>"
	"ޒަމާނީ" adj @root #4->0
"<ޑިމޮކްރެސީގެ>"
	"ޑިމޮކްރެސީ" n nhum sg def gen @nmod:poss #5->6
"<ރޫޙާއި>"
	"ރޫޙާއި" n nhum sg def soc @obj #6->4
"<ޚިލާފު>"
	"ޚިލާފު" post @advmod #7->8
"<އިސްލާޙެއް>"
	"އިސްލާޙް" n nhum sg ind dir @nmod #8->6
"<.>"
	"." sent @punct #9->4

discourse

Discourse Marker of a sentence. It's usually used to mark the modal words in a sentence. For example, it is very common in Dhivehi to have a sentence ending particle that doesn't mean anything. The discourse relation is used to mark that particle.

  • Examples:
"<އަލި>"
	"އަލި" np @nsubj #1->7
"<ލަނޑުޖަހާނެ>"
	"ލަނޑުޖަހާނެ" v tv act fut p3 @advcl #2->7
"<ވަރަށް>"
	"ވަރަށް" adv @advmod #3->4
"<ބައިވަރު>"
	"ބައިވަރު" det @det #4->5
"<ފުރުޞަތު>"
	"ފުރުޞަތު" n nhum sg def dir @obj #5->7
"<ލިބުނެވެ>"
	"އެވެ" mod @discourse #6->7
		"ލިބުން" v tv act past p3 @root #7->0
"<.>"
	"." sent @punct #8->7
"<މަދްރަސާ>"
	"މަދްރަސާ" n nhum sg def dir @nsubj #1->6
"<ގެއަށްވުރެ>"
	"ވުރެ" post @advmod #2->3
		"ގެ" n nhum sg def dat @obj #3->6
"<މާ>"
	"މާ" adv @advmod #4->6
"<ބޮޑެވެ>"
	"އެވެ" mod @discourse #5->6
		"ބޮޑު" adj @root #6->0
"<.>"
	"." sent @punct #7->6

amod

Adjective modifier of the sentence. It's usually used to mark the adjective that modifies a noun.

  • Examples:
"<ކުރަން>"
	"ކުރަ" v tv act pres p1 @fixed #1->2
"<ޖެހޭ>"
	"ޖެހޭ" v tv act pprs @root #2->0
"<ހޭދަ>"
	"ހޭދަ" adj @amod #3->4
"<އިތުރުވުން>"
	"އިތުރުވުން" n nhum sg def dir @obj #4->2
"<.>"
	"." sent @punct #5->2
"<މާދަމާ>"
	"މާދަމާ" adv @advmod #1->5
"<އަލީ>"
	"އަލީ" np @nsubj #2->5
"<ފާހަނަ>"
	"ފާހަނަ" n nhum sg def dir @obj #3->5
"<ސާފު>"
	"ސާފު" adj @amod #4->5
"<ކުރާނެ>"
	"ކުރަ" v tv act fut p3 @root #5->0
"<.>"
	"." sent @punct #6->5
Retrieved from "http://wikis.swarthmore.edu/ling073/index.php?title=Dhivehi/Universal_Dependencies&oldid=10715"