"Looksalike" algorithm

[almurph]

Hi everyone,

I'm looking around for a VB.NET algorithm that can do non-exact
matches, that is, a "looks alike" type logic for word patterns.
Does anyone have any suggestions/comments/algorthms they would like
to mention? Any comments/suggestions greatly appreciated.

Al.

 

[Patrice]

I don't remember the name right now but I've seen once someone who had a
problem on an algorithm that finds the "edit distance" between two strings
(ie. the number of elementary insertion/deletion operations your have to do
to go from one string to the other one).

Googling for "edit distance" should hopefully return pages for this. you
still may want to explain what you are trying to do. The type of application
could raise more appropriate suggestions...

 

[Patrice]

This is the Levenshtein distance :
http://en.wikipedia.org/wiki/Levenshtein_distance

If I remember the poster had a problem with comparing similarity accross
couples of words (i.e. a 20 letter word with 4 changes is still "better"
than a 4 letter word with 3 changes). I suggested to divide the edit
distance by the length of the word to get a "number of changes per
letter"...

 

[almurph]

Merci Patrice.

Al.

 

[Larry Lard]

Hi everyone,

I'm looking around for a VB.NET algorithm that can do non-exact
matches, that is, a "looks alike" type logic for word patterns.
Does anyone have any suggestions/comments/algorthms they would like
to mention? Any comments/suggestions greatly appreciated.

If you are doing fuzzy matching on names, something like Soundex might
help:

<http://en.wikipedia.org/wiki/Soundex>

Lots of stuff in there.

 

[Guest]

Hi,

In addition to the other comments you can use option compare text to
make the string comparisons non case sensitive. Like works with string
comparsions. Finally regular expressions are great for matching patterns.

http://msdn.microsoft.com/library/d...stemtextregularexpressionsregexclasstopic.asp

Ken

 

[almurph]

Thank you both for your comments