Thanks for the feedback. This is the kind of issue I anticipated: different languages work differently! Wordology already uses heuristics to try and detect related words automatically, but they're tuned for European languages. The algorithm needs more options so that the user can tune it to their particular language, and this is where I really need user feedback, because of course I don't speak every language! If you'd like to open an issue on the Github page:

https://github.com/geajack/Wordology/issues

(Click "New Issue" - you'll need to make a GitHub account)

we can discuss this at greater length.