fuzzy match % for sentences that were shortened (a part of original sentence deleted)
Thread poster: Maxime Bujakov

Maxime Bujakov  Identity Verified
France
Local time: 22:39
Member (2006)
French to English
+ ...
Jun 12

Hello!

With all great capacities of new-generation Trados studios, I have been waiting for decades now for improvements in obvious cases of fuzzy matches.
Today's one is sentences that were shortened (a part of original sentence deleted). Usually, a sentence is shortened by deleting something in parentheses or some qualifying part.

From the translator efforts point of view, this is close to 99% match (just delete the relevant piece in the target language). Let's say, 95%. And the Trados judges sometimes almost no match at all if a half-sentence is deleted.

Are you Trados guys aware of this? If so, for how many years ?


Direct link Reply with quote
 

Arianne Farah  Identity Verified
Canada
Local time: 16:39
Member (2008)
English to French
Which version are you using? Jun 12

In Studio 2017, they have the TU fragment match that you can activate which does exactly what you're asking for (you need to update your memories first though - there's an UpLift plugin for that).

Direct link Reply with quote
 

Maxime Bujakov  Identity Verified
France
Local time: 22:39
Member (2006)
French to English
+ ...
TOPIC STARTER
thank you Jun 12

Thank you, the latest Trados version I know is 2014, but plugin stuff sounds a bit complicated...
I really hope Trados people would let me know whether this is deliberate or they just never thought about making fuzzy matches more up to the point... thanks guys!


Direct link Reply with quote
 

Mirko Mainardi  Identity Verified
Italy
Local time: 22:39
Member
English to Italian
+ ...
Impact on % Jun 12

Arianne Farah wrote:

In Studio 2017, they have the TU fragment match that you can activate which does exactly what you're asking for (you need to update your memories first though - there's an UpLift plugin for that).


Does that have a direct impact on fuzzy percentages?


Direct link Reply with quote
 

Mirko Mainardi  Identity Verified
Italy
Local time: 22:39
Member
English to Italian
+ ...
Clarification Jun 12

Maxime Bujakov wrote:

Today's one is sentences that were shortened (a part of original sentence deleted). Usually, a sentence is shortened by deleting something in parentheses or some qualifying part.

From the translator efforts point of view, this is close to 99% match (just delete the relevant piece in the target language). Let's say, 95%. And the Trados judges sometimes almost no match at all if a half-sentence is deleted.


Not sure I understand you correctly. Are you saying that if you have two segments, like "The part that remains the same and the part that gets deleted" (but the order could be inverted) and "The part that remains the same", then the second one will be treated as a low match? If so, how is that a problem? Just lower the automatic lookup threshold. Additionally, sometimes it's not just as easy as in the mock sentence above, as you might have to change the modified sentence to make it sound OK by itself and adapt to the surrounding context (which might have changed as well).


Direct link Reply with quote
 

Maxime Bujakov  Identity Verified
France
Local time: 22:39
Member (2006)
French to English
+ ...
TOPIC STARTER
We have the % serving certain purposes Jun 12

Mirko Mainardi wrote:
Not sure I understand you correctly. Are you saying that if you have two segments, like "The part that remains the same and the part that gets deleted" (but the order could be inverted) and "The part that remains the same", then the second one will be treated as a low match?


Yes, I mean that.

Mirko Mainardi wrote:
If so, how is that a problem? Just lower the automatic lookup threshold.


We have the % serving certain purposes. Example: a good match is lost amid poor matches.
We also often limit some operations to high % fuzzy only. My good matches are not included then.

Trados guys, are you there?


Direct link Reply with quote
 

Arianne Farah  Identity Verified
Canada
Local time: 16:39
Member (2008)
English to French
Completely separate Jun 12

Mirko Mainardi wrote:

Arianne Farah wrote:

In Studio 2017, they have the TU fragment match that you can activate which does exactly what you're asking for (you need to update your memories first though - there's an UpLift plugin for that).


Does that have a direct impact on fuzzy percentages?


It's in a different window and has nothing to do with analysis - kind of like how the concordance search is separate from the translation results.

Basically if a sentence in the memory is ABCDEFGH and then you have a new segment that's DEFG, it's not enough of a match for the memory to pick it up, of course you would find it if you ran a concordance search, but you'd have to know what you're looking for and which strings to single out to search, Fragment matches will pull if up for you.

@Maxime - the plugin is just to update your old memories more quickly than what Studio does on its own; if you have small memories to update or start a new one from scratch, you won't need it.


Direct link Reply with quote
 

Mirko Mainardi  Identity Verified
Italy
Local time: 22:39
Member
English to Italian
+ ...
Impact on analysis Jun 12

Maxime Bujakov wrote:

We have the % serving certain purposes. Example: a good match is lost amid poor matches.
We also often limit some operations to high % fuzzy only. My good matches are not included then.


Of course, but if they did what you're suggesting (raising the fuzzy % in similar cases), that would have a direct impact on weighted wordcounts (and that would be bad news for many translators).


Direct link Reply with quote
 

Mirko Mainardi  Identity Verified
Italy
Local time: 22:39
Member
English to Italian
+ ...
Thanks Jun 12

Arianne Farah wrote:

Mirko Mainardi wrote:

Arianne Farah wrote:

In Studio 2017, they have the TU fragment match that you can activate which does exactly what you're asking for (you need to update your memories first though - there's an UpLift plugin for that).


Does that have a direct impact on fuzzy percentages?


It's in a different window and has nothing to do with analysis - kind of like how the concordance search is separate from the translation results.

Basically if a sentence in the memory is ABCDEFGH and then you have a new segment that's DEFG, it's not enough of a match for the memory to pick it up, of course you would find it if you ran a concordance search, but you'd have to know what you're looking for and which strings to single out to search, Fragment matches will pull if up for you.


Got it, thanks Arianne. I still haven't installed 2017 because of the many issues it had after launch, but this sounds like an interesting feature.


Direct link Reply with quote
 

Soonthon LUPKITARO(Ph.D.)  Identity Verified
Thailand
Local time: 04:39
Member (2004)
English to Thai
+ ...
Definition of % Jun 12

The fuzzy matches % in Trados is an algorithm to calculate statistical hits of segments in question.
I just read the % only as guideline. Human judgment is essential to use this CAT efficiently.
If I cannot enjoy its %, I can recalculate the segment hits by using external algorithm e.g. codes in Visual Basic Language.

Dr. Soonthon Lupkitaro
Bangkok


Direct link Reply with quote
 

Maxime Bujakov  Identity Verified
France
Local time: 22:39
Member (2006)
French to English
+ ...
TOPIC STARTER
We have the % serving certain purposes Jun 13

Soonthon LUPKITARO(Ph.D.) wrote:

The fuzzy matches % in Trados is an algorithm to calculate statistical hits of segments in question.
I just read the % only as guideline. Human judgment is essential to use this CAT efficiently.
If I cannot enjoy its %, I can recalculate the segment hits by using external algorithm e.g. codes in Visual Basic Language.

Dr. Soonthon Lupkitaro
Bangkok


thank you but I already answered:

We have the % serving certain purposes. Example: a good match is lost amid poor matches.
We also often limit some operations to high % fuzzy only. My good matches are not included then.


Direct link Reply with quote
 

Maxime Bujakov  Identity Verified
France
Local time: 22:39
Member (2006)
French to English
+ ...
TOPIC STARTER
that would be bad news for many translators Jun 13

Mirko Mainardi wrote:

Maxime Bujakov wrote:

We have the % serving certain purposes. Example: a good match is lost amid poor matches.
We also often limit some operations to high % fuzzy only. My good matches are not included then.


Of course, but if they did what you're suggesting (raising the fuzzy % in similar cases), that would have a direct impact on weighted wordcounts (and that would be bad news for many translators).


I represent an agency

that said, I do not practice fuzzy reduction


Direct link Reply with quote
 


To report site rules violations or get help, contact a site moderator:


You can also contact site staff by submitting a support request »

fuzzy match % for sentences that were shortened (a part of original sentence deleted)

Advanced search







TM-Town
Manage your TMs and Terms ... and boost your translation business

Are you ready for something fresh in the industry? TM-Town is a unique new site for you -- the freelance translator -- to store, manage and share translation memories (TMs) and glossaries...and potentially meet new clients on the basis of your prior work.

More info »
SDL Trados Studio 2017 Freelance
The leading translation software used by over 250,000 translators.

SDL Trados Studio 2017 helps translators increase translation productivity whilst ensuring quality. Combining translation memory, terminology management and machine translation in one simple and easy-to-use environment.

More info »



Forums
  • All of ProZ.com
  • Term search
  • Jobs
  • Forums
  • Multiple search