Fuzzy matching in term bases
Posted by Péter Botta on 21 May 2013 09:00 AM
Title: Fuzzy matching in term bases
Description: In memoQ 2013, the fuzzy term lookup is introduced. You do not need to migrate your term bases from memoQ 6.2.
create a new term base in memoQ 2013, you have a new option in your term base
defaults: Fuzzy. Fuzzy means to get hits independently of prefix or suffix
matching. In case of German, for instance, you have "Mutter" in your
term base. With the Fuzzy setting, it also finds "Mütter". German has
a lot of composita like "Festplattenspeicherplatz". When your term
base contains "Festplatte" and "speicher", it will be highlighted
in the source cell containing the word "Festplattenspeicherplatz".
The internal threshold for finding fuzzy matches is 80% similarity; if below, you do not get a fuzzy match. For example in German: Erfüllungsaufwand is singular, the fuzzy matching also finds Erfüllungsaufwände which is plural, but it does not find Erfüllungsaufwendungen.
You also get hits in the translation grid for terms which may have a typo in the term base. For instance, you have entered the term “Stuhl” in your term base, but this is a typo, you actually meant to enter the term “Stahl”. With the Fuzzy matching setting in your term base, you will get a term base hit in the translation grid, although the term has a typo.