This should point you in the right direction I think:
The levenshtein algorithm basically gives you the “edit-distance”
between two strings. E.g. the minimum amount of
insertions/replacements/deletions to make the strings identical. It
gives you a pretty good indication on how similar the strings are.
Soundex transforms all strings that are similar into the same 4
character code (which looks something like “E246”).
Metaphone is preferred over soundex I believe. It also transforms
similar strings into the same character sequence, but doesn’t limit
itself to just 4 characters. That means it works a bit better with
There are probably other algorithms around as well, but I’ve had pretty
good luck with these three.