You want to find similar words to one you may be thinking of against a dictionary in your computer. How to do so?1- Get only those words whith a Levenshtein's distance which migth seem acceptable to you.
2- Convert both the dictionary words and the original one to a binary array. F.e.: 'AAA' -> '010000010100000101000001'
3- Get the crossed-correlation coeficient between the oroginal and the dictionary words.
4- Print only those whose crossed-correlation coeficient at bit-level is higher than 94% (or whatever you want):
5- Amazing eh? I didn't either thought it could work so well (at least it does in spanish)
6- Carefull!-> It's slow an uses a lot of CPU.
Printer-friendly copy
Email this topic to a friend
