Termbase match preferences not working

I'm having the following problem with SDL Trados. When translating Chinese>English, if the term match threshhold is below 90% or so, the termbase will produce incorrect matches and errors. I change the setting from the default of 70% to a 90%, but after each project it resets and I have to change it again, no matter where I change it, whether in the preferences screen or elsewhere. If not changed, then in basically every project, I will be bombarded with extremely large numbers of incorrect hits. 

  • Hi  

    I think this is going to be linked to the difficulties in tokenizing this language and there isn't an easy language. We are hoping to address this with smarter ways to handle terminology in the future and we have started to do this with the new Language Cloud Terminology solution you can find in SR1.  So here you have an additional option that we are working on:

    This search type uses stemming technology in the background and this should help improve the search capability for many languages.  This initial release supports 19 languages but unfortunately Chinese isn't one of them.  I'm not sure of the roadmap for additional languages yet but I thought it would be interesting at least for you to know that we are investigating ways to improve the features in the products.

    So to summarise I don't think there is a simple solution for you here other than to make sure the threshold is maintained at a higher level.

    Regards

    Paul

    Paul Filkin | RWS Group

    ________________________
    Design your own training!

    You've done the courses and still need to go a little further, or still not clear? 
    Tell us what you need in our Community Solutions Hub

  • Hi Paul! Thanks again for the very interesting answer. I see we just got match repair for Chinese, which seems like it will be very promising, and I look forward to seeing some new features for Chinese! When translating from Chinese as a source language, since virtually everything is a 2-character expression, locking it to 90%+ usually does the trick. The term checkers have gotten confused however where the same Chinese word can be translated to every different kind of tense and form of a word. (i.e. the same word count be, devalue, devaluation, devaluing, devalues, devalued etc., or could be inoperable / not operable)

    In that case, is there a way to make the threshold default to 90% so that I don't have to change it every time? No matter what I do, the setting always goes back to 70%.
  • Hi Adrian,

    I think that if you have your termbase search threshold already setup in your Global Options then this setting should apply for all new projects. At least it worked this way for me when I tested it just now. Can you explain what you are doing to try and achieve this?

    Regards

    Paul

    Paul Filkin | RWS Group

    ________________________
    Design your own training!

    You've done the courses and still need to go a little further, or still not clear? 
    Tell us what you need in our Community Solutions Hub

  • I might be doing something wrong since I'm still getting the reset to the 70% threshold. The steps I just took were to go to the Welcome tab -> options -> all language pairs -> confirm threshold at 90%. Then I went to create new project -> existing template -> new project editor -> project options, and the threshold is at 70%. However, in File>Options>All Language pairs, I see the threshold at 90%. Any thoughts on what I'm doing wrong?