|
First of all you need a data structure of every word in the language(s) your site is using.
You can of course limit this to every word on your site (maybe 3 letters and over) to reduce storage space.
Also to save even more space you could 'tokenise' combinations of letters (this is how spell checkers compress the data on your wordprocessor spellcheck/dictionary)
Then you need a lookup structure of mis-spellings and synonyms etc, this can of course be "self-learning" insofar as the user selecting a replacement also adds to the structure.
Your system first searches the primary word structure and if no matches are found, searches the lookup structure and offers the choice(s) to the users.
This is obviously a simplistic overview and the reality is somewhat more complex.
__________________
Chris. ->> Please login or register to view this content. Registration is FREE <<-
A foolish consistency is the hobgoblin of little minds
Thought for today:- Is SEO the only industry where all the cowboys are Indians?
|