Thread Rating:
  • 1 Vote(s) - 5 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Digraph Compression
britlion Wrote:
boriel Wrote:
britlion Wrote:Interesting. What have you used them for, Boriel?
I create my own statistical language recognizer, by previously analysing statistical frequency of digraphs and trigraphs. Then for a new text, I do the same, and return the "guessed" language by taken the closest histogram to this new one. :roll:

Ooh. I'd have never thought of that one. I'd have been messing with dictionaries and highest count of recognised words, I think.

I wonder how google translate does it.
I used "trigraphs" later,and works much better. For 10+ words, the accuracy I got is near 100%.
To "train" the frequencies I use these word count lists from here.

Messages In This Thread

Forum Jump:

Users browsing this thread: 1 Guest(s)