Difference between revisions of "Turkic-Turkic translator"

Revision as of 19:47, 22 September 2013

Pair	Languages	Location	Directions	Testvoc	Coverage (%)	WER	WER*
apertium-kaz-tat	Kazakh — Tatar	trunk	`kaz-tat`, `tat-kaz`	—	~88.7%^(kaz), ~87%^(tat)	-	-
apertium-tr-az	Turkish — Azeri	trunk	`tr-az`	-	-	-	-
apertium-tur-kir	Turkish — Kyrgyz	staging	`tur-kir`	-	(`tr`: 80.39)	-	-
apertium-tt-ba	Tatar — Bashkir	nursery	`tt-ba`, `ba-tt`	-	-	-	-
apertium-tuk-tur	Turkmen — Turkish	nursery	`tuk-tur`, `tur-tuk`	-	-	-	-
apertium-tr-tt	Turkish — Tatar	incubator	`tr-tt`, `tt-tr`	-	-	-	-
apertium-cv-tr	Chuvash — Turkish	incubator	`tr-cv`, `cv-tr`	-	-	-	-
apertium-kaz-kir	Kazakh — Kyrgyz	incubator	`kaz-kir`	—	~87.3%^(kaz), ~85.8%^(kir)	—	—
apertium-tur-uzb	Turkish - Uzbek	incubator	`uzb-tur`	—	—	—	—

WER is the idealised word error rate with full coverage on a crafted test set.
WER* is the realistic error rate given non-complete coverage.
Coverage is the naïve coverage over one or more free corpora.

This is a general workplan for development efforts for Turkic--Turkic prototype translators. Suitable for use in the Google Summer of Code.

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 $n$ words and performing an evaluation for post-edition word error rate (WER). The output for those $n$ words should be clean.

Week	Dates	Trimmed coverage	Testvoc	Evaluation	Notes
0	23/04—21/05	45%	`<postadv>` `<ij>`	500 words	Preliminary evaluation. Translate the story total coverage and without diagnostics (in both directions). Get a baseline WER. Work on disambiguation, the morphological ambiguities in the story should be resolved.
1	21/05—27/05	50%	`<num>` `<post>`	-	Basic numerals, and postpositions should be clean.
2	28/05—03/06	53%	`<cnjcoo>` `<cnjadv>` `<cnjsub>`	-
3	04/06—10/06	59%	`<adv>`	200 words	Work on disambiguation.
4	11/06—17/06	63%	`<prn>` `<det>`	-
5	18/06—24/06	68%	`<adj>` `<adj><advl>`	-	Adjectives with attributive and adverbial function should be clean. Numerals + affixes should have their phonology solved.
6	25/06—01/07	70%	`<n>` `<num><subst>` `<np>` `<adj><subst>`	500 words	Midterm evaluation. Work on disambiguation.
7	02/07—08/07	73%	-	-
8	09/07—15/07	75%	-	-
9	16/07—22/07	77%	-	200 words	Work on disambiguation.
10	23/07—29/07	80%	`<v>`	-
11	30/07—05/08	80%	-	-	Work on disambiguation.
12	06/08—12/08	80%	all categories clean	500 words	Final evaluation. Tidying up, releasing

@@ Line 22: / Line 22: @@
 | [[apertium-kaz-kir]] || [[Kazakh]] — [[Kyrgyz]] || [[incubator]] || <code>kaz-kir</code> || — || ~{{:Apertium-kaz-kir/stats/kaz-average}}%{{slc|kaz}}, ~{{:Apertium-kaz-kir/stats/kir-average}}%{{slc|kir}} || — || —
 |-
-| [[apertium-tur-uzb]] || [[Turkish]] - [[Uzbek]] || [[incbuator]] || <code>uzb-tur</code> || — || — || — || —
+| [[apertium-tur-uzb]] || [[Turkish]] - [[Uzbek]] || [[incubator]] || <code>uzb-tur</code> || — || — || — || —
 |}