Difference between revisions of "Biak/Disambiguation"

From LING073
Jump to: navigation, search
(Initial Evaluation of Ambiguity)
 
(16 intermediate revisions by 2 users not shown)
Line 1: Line 1:
 +
==Our Code==
 +
[https://github.swarthmore.edu/Ling073-sp21/ling073-bhw/blob/master/apertium-bhw.bhw.rlx The .rlx file on our Github]
 +
 
==Initial Evaluation of Ambiguity==
 
==Initial Evaluation of Ambiguity==
* Our initial ambiguity = 1.06 = 4169/3932
+
* Our initial ambiguity = 1.012
 
* Some examples:
 
* Some examples:
** mran{{tag|v}}{{tag|iv}}{{tag|p1}}{{tag|pl}}{{tag|inc}} ra{{tag|pr}}/ra{{tag|cnjsub}} vanda{{tag|n}} -> komran ra vanda -> We walk ''along'' the seaside
+
** {{morphTest|mran{{tag|v}}{{tag|iv}}{{tag|p1}}{{tag|pl}}{{tag|inc}} ra{{tag|pr}}/ra{{tag|cnjsub}} vanda{{tag|n}}|komran ra vanda}} # We walk ''along'' the seaside
 +
** {{morphTest|saroi{{tag|n}}+i{{tag|prn}}{{tag|pers}}{{tag|p3}}{{tag|sg}}{{tag|spc}} orn{{tag|v}}{{tag|tv}}{{tag|p3}}{{tag|sg}} si{{tag|prn}}{{tag|pers}}{{tag|p3}}{{tag|pl}}{{tag|an}}{{tag|spc}}+ya{{tag|det}}{{tag|dem}} inm{{tag|v}}{{tag|tv}}{{tag|p3}}{{tag|an}} monda{{tag|adv}} ra{{tag|pr}}/ra{{tag|cnjsub}} mar{{tag|v}}{{tag|iv}}{{tag|p3}}{{tag|pl}}{{tag|an}}|saroiya dorn siya, siya sinm monda ra simar}} # The whale ate some, others drank (river water) only ''until'' they died.
  
 
==Final Evaluation of Ambiguity==
 
==Final Evaluation of Ambiguity==
 +
* Final ambiguity = 1.010
 +
 +
* We created a rule which selects the conjunction if the next word is a verb, in which case it cannot be a preposition. This removed the ambiguity described in the examples above.
 +
* We found ambiguity between the possessive inalienable noun 'nari' (tooth 1/3sg) and 'nari' meaning will, to indicate some sort of future tense.
 +
** There is a lack of examples and documentation on these forms, particularly the tense form, which was unmentioned in our comprehensive grammar. We have yet to determine a steadfast rule to disambiguate these words grammatically.
 +
** We were able to cover some cases, where we remove the auxiliary 'will' if there are no more verbs in the sentence. This will be helpful when the noun is being used as a subject.
 +
* Ambiguity should be able to be lowered by increasing corpus coverage (in particular, adding more verb stems).
  
[[Category:Sp21_Disambiguation]]
+
[[Category:Sp21_Disambiguation]][[Category:Biak]]

Latest revision as of 14:30, 4 April 2021

Our Code

The .rlx file on our Github

Initial Evaluation of Ambiguity

  • Our initial ambiguity = 1.012
  • Some examples:
    • mran<v><iv><p1><pl><inc> ra<pr>/ra<cnjsub> vanda<n> ↔ komran ra vanda # We walk along the seaside
    • saroi<n>+i<prn><pers><p3><sg><spc> orn<v><tv><p3><sg> si<prn><pers><p3><pl><an><spc>+ya<det><dem> inm<v><tv><p3><an> monda<adv> ra<pr>/ra<cnjsub> mar<v><iv><p3><pl><an> ↔ saroiya dorn siya, siya sinm monda ra simar # The whale ate some, others drank (river water) only until they died.

Final Evaluation of Ambiguity

  • Final ambiguity = 1.010
  • We created a rule which selects the conjunction if the next word is a verb, in which case it cannot be a preposition. This removed the ambiguity described in the examples above.
  • We found ambiguity between the possessive inalienable noun 'nari' (tooth 1/3sg) and 'nari' meaning will, to indicate some sort of future tense.
    • There is a lack of examples and documentation on these forms, particularly the tense form, which was unmentioned in our comprehensive grammar. We have yet to determine a steadfast rule to disambiguate these words grammatically.
    • We were able to cover some cases, where we remove the auxiliary 'will' if there are no more verbs in the sentence. This will be helpful when the noun is being used as a subject.
  • Ambiguity should be able to be lowered by increasing corpus coverage (in particular, adding more verb stems).