Hi Spiros Doikas, we made changes recently that should actually have enhanced all cases involving what we call "acronyms" which you have here - i.e. one acronym segments. I agree this is not optimal, I am wondering if we should retrieve a fuzzy match at all in such cases. I will try to find out more details and will hopefully be able to get back to you in this posting here. BTW - are you using file-based or server/GroupShare-based TMs? Thanks, Daniel

Hi Daniel, file-based.

Thanks Spiros Doikas. I have checked and was jumping to incorrect assumptions above. Basically the optimisations for acronym handling are not yet fully available. As you are below 70%, you will retrieve more "noise" from the TM in such cases, and since what we see as acronyms (all upper case word), we are trying to be maybe too smart in retrieving fuzzy matches which are really no matches in such cases. I would say let's watch this space to see what future enhancements to this might bring. Thanks

Thanks Spiros Doikas. I have checked and was jumping to incorrect assumptions above. Basically the optimisations for acronym handling are not yet fully available. As you are below 70%, you will retrieve more "noise" from the TM in such cases, and since what we see as acronyms (all upper case word), we are trying to be maybe too smart in retrieving fuzzy matches which are really no matches in such cases. I would say let's watch this space to see what future enhancements to this might bring. Thanks

Here is another issue on the definition of concepts: not all capitalized words are acronyms :)

... which is why you can disable acronym recognition in such cases in the TM altogether, this will then treat them as normal words.

Agreed; but an application should be able to be more intuitive out-of-the-box in handling such scenarios instead of outputting completely irrelevant matches.