Difference between revisions of "Apertium-kaz-kir/stats"

From Apertium
Jump to navigation Jump to search
(→‎kaz: moar 70278)
(→‎kir: r70278)
Line 99: Line 99:
 
* words: <section begin=azattyk2010-words />3.4M<section end=azattyk2010-words />
 
* words: <section begin=azattyk2010-words />3.4M<section end=azattyk2010-words />
 
* regular coverage: ~{{#lst:Apertium-kir/stats|azattyk2010-coverage}}%
 
* regular coverage: ~{{#lst:Apertium-kir/stats|azattyk2010-coverage}}%
* trimmer coverage: ~<section begin=azattyk2010-coverage />82.6<section end=azattyk2010-coverage />%
+
* trimmer coverage: ~<section begin=azattyk2010-coverage />89.2<section end=azattyk2010-coverage />%
* as of: r47474
+
* as of: r70278
   
 
azattyk2009
 
azattyk2009
 
* words: <section begin=azattyk2009-words />4.1M<section end=azattyk2009-words />
 
* words: <section begin=azattyk2009-words />4.1M<section end=azattyk2009-words />
 
* regular coverage: ~{{#lst:Apertium-kir/stats|azattyk2009-coverage}}%
 
* regular coverage: ~{{#lst:Apertium-kir/stats|azattyk2009-coverage}}%
* trimmer coverage: ~<section begin=azattyk2009-coverage />82.0<section end=azattyk2009-coverage />%
+
* trimmer coverage: ~<section begin=azattyk2009-coverage />86.25<section end=azattyk2009-coverage />%
* as of: r47474
+
* as of: r70278
   
 
bible
 
bible
 
* words: <section begin=bible-words />174K<section end=bible-words />
 
* words: <section begin=bible-words />174K<section end=bible-words />
 
* regular coverage: ~{{#lst:Apertium-kir/stats|bible-coverage}}%
 
* regular coverage: ~{{#lst:Apertium-kir/stats|bible-coverage}}%
* trimmer coverage: ~<section begin=kirbible-coverage />82.5<section end=kirbible-coverage />%
+
* trimmer coverage: ~<section begin=kirbible-coverage />86.01<section end=kirbible-coverage />%
* as of: r47474
+
* as of: r70278
   
 
wp2011/04
 
wp2011/04
 
* words: <section begin=wp2011/04-words />545K<section end=wp2011/04-words />
 
* words: <section begin=wp2011/04-words />545K<section end=wp2011/04-words />
 
* regular coverage: ~{{#lst:Apertium-kir/stats|wp2011/04-coverage}}%
 
* regular coverage: ~{{#lst:Apertium-kir/stats|wp2011/04-coverage}}%
* trimmer coverage: ~<section begin=wp2011/04-coverage />75.5<section end=wp2011/04-coverage />%
+
* trimmer coverage: ~<section begin=wp2011/04-coverage />78.92<section end=wp2011/04-coverage />%
* as of: r47474
+
* as of: r70278
   
 
== WER ==
 
== WER ==

Revision as of 04:38, 22 June 2016

Over-all stats

apertium-kaz

  • trimmed stems as of: r46670
  • stats:
full trimmed
coverage ~94.5% ~87.3%
stems 36,595 6512


apertium-kir

  • trimmed stems as of: r46670
  • stats:
full trimmed
coverage ~90.4% ~85.8%
stems 14,424 6532

bidix

  • as of: r49573

Trimmed coverage

kaz

Әуезов

  • words: 155K
  • regular coverage: ~92.89%
  • trimmed coverage: ~85.51%
  • as of: r70278
  • wikipage: Әуезов corpus

bible

  • words: 577K
  • regular coverage: ~95.29%
  • trimmed coverage: ~88.9%
  • as of: r70278

azattyq2010

  • words: 3.2M
  • regular coverage: ~95.07%
  • trimmed coverage: ~89.52%
  • as of: r70278
  • wikipage: RFERL_corpora

azattyq2012

  • words: 3.9M
  • regular coverage: ~%
  • trimmed coverage: ~89.20%
  • as of: r70278
  • wikipage: RFERL_corpora

wp2011

  • words: 850K
  • regular coverage: ~90.72%
  • trimmed coverage: ~80.1%
  • as of: r47474

wp100K

  • words: 1.7M
  • regular coverage: ~%
  • trimmed coverage: ~82.61%
  • as of: r70274

quran

  • words: 107K
  • regular coverage: ~96.71%
  • trimmed coverage: ~92.49%
  • as of: r70278

kir

azattyk2010

  • words: 3.4M
  • regular coverage: ~92.11%
  • trimmer coverage: ~89.2%
  • as of: r70278

azattyk2009

  • words: 4.1M
  • regular coverage: ~92.04%
  • trimmer coverage: ~86.25%
  • as of: r70278

bible

  • words: 174K
  • regular coverage: ~92.25%
  • trimmer coverage: ~86.01%
  • as of: r70278

wp2011/04

  • words: 545K
  • regular coverage: ~85.37%
  • trimmer coverage: ~78.92%
  • as of: r70278

WER

kaz→kir

trt_2013-04-05

  • words: 245
  • WER: 6.12%
  • PER: 4.90%
  • as of: r44975
  • type: training

story

  • words: 351
  • WER: 4.84%
  • PER: 3.99%
  • as of: r44975
  • type: training