Difference between revisions of "Bikol/Disambiguation"

Revision as of 01:00, 1 March 2017

Possible ambiguities:

• asin (and or salt)
• kan (of or can)
• tapos (end or then) / 227 304
  • Patoson mo an pansit, tápos, darahon mo digdí. (230)
  • Tapos na an síne? Tapos ka na? Tapos mo nang gamiton an lapis? (304)
• diyan (later or place)
• ta (because/ ni class inclusive pronoun) / 59 66
  • Tano ta inapod mo sinda? Inapod ko sinda ta mahali na kami. (59)
  • Inapód ta si Márilyn? Iyó, inapód ta siyá. / Inapod ta si Daisy? Da'í ta siyá inapód.(66)

Personal note:

Use the following command to find all disambiguities in the corpus.

cat ~/Documents/ling073-bcl-corpus/bcl.corpus.basic.txt | lt-proc bcl.automorf.bin | sed 's/$\W*\^/$\n^/g' | grep "\/.*\/"

Initial evaluation of ambiguity

• Number of tokens: 1967
• Number of analyses: 2059
• Current ambiguity: 1.04677 analyses/token

However, none of these are actually ambiguous cases, but rather represent tokenization errors. So, the effective ambiguity is 1.

After the addition of two example sentences illustrating the two meanings of "tápos" (explained below), here are the ambiguity statistics:

• Number of tokens: 1983
• Number of analyses: 2079
• Current ambiguity: 1.04841 analyses/token

Tápos

The Bikol word "tápos" can have either of the following two meanings, illustrated by example sentences pulled from Bikol Text by Malcolm W. Mintz. These sentences were added to my corpus.

• Conjunctive adverb: then, afterward
  • Patosón mo an pansít, tápos, darahón mo digdí. (Wrap the noodles, then bring them here.)
• Adjective: over, completed
  • Tápos na an síne. (The movie is already over.)

Final evaluation of ambiguity

• Number of tokens: 1983
• Number of analyses: 2075
• Current ambiguity: 1.04639 analyses/token

The number of analyses was reduced by 4 after disambiguation. This corresponds with the four instances of "tápos" in my corpus (two originally in my corpus and the two from the example sentences which I added). This indicates that all four sentences containing "tápos" in my corpus were successfully disambiguated.