This article is about the implementation of the new analysis system - a more transparent, and clearer repartition of the volume based on correlations, existing matches, and anticipated matches.


Affected users

  • Any role that have access to the Analysis panel
  • Translator, Reviewer, Manager, Customer, SEO, Legal, Final


Affected versions

  • Wezen version 11 and above


Related articles


New Analysis terminologies

Wezen's analysis update introduced three categories of volume repartition: correlations, existing matches, and new, in which there are three sub-categories of new: anticipated matches, machine translated, and strict new.




Correlations

There are two categories of correlations repartition.


  1. Existing correlations
    • Exact matches with previous translated segments within the correlation scope
  2. Anticipated correlations
    • Exact matches with previous untranslated segments within the correlation scope


Depending on the repetition scope of your project, they will have different definitions. Below is a table to summarize them:

Repetition scope type
NoneTaskResubmissionProjectFragment
Existing correlations--Correlations with tasks with same external ID, languages, and translated source segmentCorrelations with translated, exact matches for source segments from within the project, in the same languages

Correlations with translation units (or fragments) with the same external name, languages, and translated source content

Anticipated correlations-In-task repetitions*Correlations with tasks with same external ID, languages, and untranslated source segment + in-task repetitions*Correlations with untranslated, exact matches for source segments from within the project, in the same languages + in-task repetitions*

Correlations with translation units (or fragments) with the same external name, languages, and untranslated source content

*In-task repetitions - Repeated source segments will be correlated automatically with its parent segment(s) within the task


Existing matches

There are five categories of existing matches repartition.


  1. Existing 100%
    • Exact matches with existing segments in the TM outside the correlation scope
  2. Existing 95-99%
  3. Existing 85-94%
  4. Existing 75-84%
  5. Existing 50-74%
    • Partial matches with an existing segment in the TM


These categories repartition depends on the memory entries on the memory at the state, when the tasks undergo through the Analysis step.


New

There are three sub-categories that consist of the new words repartition.


  1. Anticipated matches - which consists of 5 categories
    1. Anticipated 100%
      • Exact matches with untranslated segments outside the correlation scope
    2. Anticipated 95-99%
    3. Anticipated 85-94%
    4. Anticipated 75-84%
    5. Anticipated 50-74%
      • Partial matches with untranslated segments outside the correlation scope
  2. Machine translated
  3. Strict New
    • Segments below the match percentage of the autofill threshold. 


By default, the sum of your general new words is the sum of all anticipated matches, machine translated, and strict new words.


This is shown on the last row of the analysis panel (highlighted in green)



You can access the detailed version of the New words repartition by clicking on the dropdown arrow (right next to New)