Difference between revisions of "Indonesian and Malaysian/Work plan"

From Apertium
Jump to navigation Jump to search
Line 6: Line 6:
 
# Extracting bilingual dictionaries from parallel corpus. Tutorial is [http://wiki.apertium.org/wiki/Extracting_bilingual_dictionaries_with_Giza%2B%2B here].
  
# Extracting bilingual dictionaries from parallel corpus. Tutorial is [http://wiki.apertium.org/wiki/Extracting_bilingual_dictionaries_with_Giza%2B%2B here].
 
#*[http://opus.lingfil.uu.se/ Open Parallel Corpus] (needs to be preprocessed first)
  
#*[http://opus.lingfil.uu.se/ Open Parallel Corpus] (needs to be preprocessed first)
  +
==Work plan==
−
==Community bonding period (April 24 - May 21)==
  
−
===Todo list===
  
−
# <s>Convert the Malaysian dictionary to Apertium format</s>
  
−
# <s>Create a script to get Indonesian word list</s>
  
−
# <s>Adding missing words from the story</s>
  
−
# <s>Adding conjunctives and interjections</s>
  
   
  +
This is a draft workplan for development of the Serbo-Croatian and Slovenian translator.
−
===Resources for the word list===
  
  +
−
* [http://pusatbahasa.kemdiknas.go.id/kbbi/ KBBI Daring]
  
  +
* Trimmed coverage means the coverage of both of the morphological analysers after being trimmed according to the bilingual dictionary of the pair, that is, only containing stems which are also in the bilingual dictionary.
−
* [http://dictionary.bhanot.net/index.html Dr Bhanot's Malay-English Dictionary] (10,000 words)
  
  +
* Testvoc for a category means that the category is [[testvoc]] clean, in both translation directions.
  +
* Evaluation is taking <math>n</math> words and performing an [[evaluation]] for post-edition word error rate (WER). The output for those <math>n</math> words should be clean.
  +
  +
{|class=wikitable
  +
|-
  +
! Week !! Dates !! Trimmed coverage !! Testvoc !! Evaluation !! Notes
  +
|-
  +
| 0 || 23/04&mdash;21/05 || 45% || {{tag|ij}} {{tag|cnjcoo}} {{tag|cnjsub}} {{tag|cnjadv}} || 500 words || '''Preliminary evaluation'''. Translate the story total coverage and without diagnostics. Get a baseline WER.
  +
|-
  +
| 1 || 21/05&mdash;27/05 || 50% || - || - ||
  +
|-
  +
| 2 || 28/05&mdash;03/06 || 53% || - || - ||
  +
|-
  +
| 3 || 04/06&mdash;10/06 || 59% || - || 200 words ||
  +
|-
  +
| 4 || 11/06&mdash;17/06 || 63% || - || - ||
  +
|-
  +
| 5 || 18/06&mdash;24/06 || 68% || - || - ||
  +
|-
  +
| 6 || 25/06&mdash;01/07 || 70% || - || 500 words || '''Midterm evaluation'''.
  +
|-
  +
| 7 || 02/07&mdash;08/07 || 73% || - || - ||
  +
|-
  +
| 8 || 09/07&mdash;15/07 || 75% || - || - ||
  +
|-
  +
| 9 || 16/07&mdash;22/07 || 77% || - || 200 words ||
  +
|-
  +
| 10 || 23/07&mdash;29/07 || 80% || - || - ||
  +
|-
  +
| 11 || 30/07&mdash;05/08 || 80% || - || - ||
  +
|-
  +
| 12 || 06/08&mdash;12/08 || 80% || ''all categories clean'' || 500 words || '''Final evaluation'''. Tidying up, releasing
  +
|-
  +
|}
   
 
[[Category:Indonesian and Malaysian|*]]
  
[[Category:Indonesian and Malaysian|*]]

Revision as of 08:47, 13 May 2012

Ideas for getting Indonesian-Malaysian bilingual dictionaries

  1. Filtering the Indonesian lemma list.
    • Get a large number of Indonesian lemmata from kateglo.
    • For each lemma, check whether they are also valid Malaysian words at PRPM.
  2. Interlanguage wiki links. Tutorial is here.
  3. Extracting bilingual dictionaries from parallel corpus. Tutorial is here.

Work plan

This is a draft workplan for development of the Serbo-Croatian and Slovenian translator.

  • Trimmed coverage means the coverage of both of the morphological analysers after being trimmed according to the bilingual dictionary of the pair, that is, only containing stems which are also in the bilingual dictionary.
  • Testvoc for a category means that the category is testvoc clean, in both translation directions.
  • Evaluation is taking words and performing an evaluation for post-edition word error rate (WER). The output for those words should be clean.
Week Dates Trimmed coverage Testvoc Evaluation Notes
0 23/04—21/05 45% <ij> <cnjcoo> <cnjsub> <cnjadv> 500 words Preliminary evaluation. Translate the story total coverage and without diagnostics. Get a baseline WER.
1 21/05—27/05 50% - -
2 28/05—03/06 53% - -
3 04/06—10/06 59% - 200 words
4 11/06—17/06 63% - -
5 18/06—24/06 68% - -
6 25/06—01/07 70% - 500 words Midterm evaluation.
7 02/07—08/07 73% - -
8 09/07—15/07 75% - -
9 16/07—22/07 77% - 200 words
10 23/07—29/07 80% - -
11 30/07—05/08 80% - -
12 06/08—12/08 80% all categories clean 500 words Final evaluation. Tidying up, releasing
Retrieved from "https://wiki.apertium.org/w/index.php?title=Indonesian_and_Malaysian/Work_plan&oldid=34203"