Project Management

Translation Memory Threshold Optimizer (TMS)

Content is machine translated from English by Phrase Language AI.

Available for

  • Team, Professional, Business and Enterprise plans

Get in touch with Sales for licensing questions.

Available for

  • Team, Ultimate and Enterprise plans (Legacy)

Get in touch with Sales for licensing questions.

To set up optimal TM thresholds and save editing time, users are offered recommended TM threshold values.

The recommended TM threshold value indicates the TM match at (or above) which it becomes faster to edit the TM pre-translated segments than the MT pre-translated segments (shorter editing time).

Note

Due to continuous improvements, the user interface may not be exactly the same as presented in the video.

TM threshold visibility

TM threshold value recommendations are provided for organizations meeting the following criteria (jointly):

  • Have jobs completed within the previous 6 months.

  • Have at least 1000 MT pre-translated segments.

  • Have at least one fuzzy score bucket (70-74%, 75-79%, 80-84%, 85-89%, 90-94%, 95-99%) with 300 or more TM pre-translated segments.

For organizations that don’t meet the above criteria, the recommended TM threshold is an average value aggregated from all Phrase TMS anonymized data.

Locations within Phrase TMS that TM thresholds are viewed:

Where

Customer settings

Recommended TM threshold

Organization settings

Following the default TM threshold

New suggested TM threshold pre-populated in the edit field; users can still modify it and set to a desired value

Organization settings

TM threshold configured manually but lower than the recommended value

Existing threshold untouched in the edit field; recommended new threshold suggested in the frame below

Organization settings

TM threshold equal or higher than the recommended value

No suggestion displayed

Project settings

TM threshold below the recommended value

Existing threshold untouched in the edit field; recommended new threshold suggested in the frame below

Project settings

TM threshold equal or higher than the recommended value

No suggestion displayed

Data used for the calculation of the recommended TM thresholds are viewed in Phrase Analytics.

TM threshold calculation

An example calculation:

MT segments

avg editing time (s)

4350

7.801

  1. Gather data from previous 6 months.

    TM fuzzy score bucket

    TM segments

    avg editing time (s)

    0-69%

    2543

    9.36

    70-74%

    375

    9.23

    75-79%

    376

    8.96

    80-84%

    423

    8.39

    85-89%

    383

    8.56

    90-94%

    155

    8.55

    95-99%

    2057

    7.799

    100-101%

    9416

    1.98

  2. Remove unrecommended fuzzy scores (below 70, over 100).

    TM fuzzy score bucket

    TM segments

    avg editing time (s)

    70-74%

    375

    9.23

    75-79%

    376

    8.96

    80-84%

    423

    8.39

    85-89%

    383

    8.56

    90-94%

    155

    8.55

    95-99%

    2057

    7.799

  3. Remove data points where there are not enough segments (1000 for MT, 300 for TM).

    TM fuzzy score bucket

    TM segments

    avg editing time (s)

    70-74%

    375

    9.23

    75-79%

    376

    8.96

    80-84%

    423

    8.39

    85-89%

    383

    8.56

    95-99%

    2057

    7.799

  4. Remove fuzzy scores that have a lower editing time than MT.

    TM fuzzy score bucket

    TM segments

    avg editing time (s)

    70-74%

    375

    9.23

    75-79%

    376

    8.96

    80-84%

    423

    8.39

    85-89%

    383

    8.56

  5. Select the highest fuzzy score remaining from step 4 (85-89%) and set that as the recommended threshold (87%)

Was this article helpful?

Sorry about that! In what way was it not helpful?

The article didn’t address my problem.
I couldn’t understand the article.
The feature doesn’t do what I need.
Other reason.

Note that feedback is provided anonymously so we aren't able to reply to questions.
If you'd like to ask a question, submit a request to our Support team.
Thank you for your feedback.