Routine with fuzzy logic to determine the relative comparison of two strings?

  • Thread starter Thread starter Elmer Smurdley
  • Start date Start date
E

Elmer Smurdley

Anyone have a canned subroutine that can determine how close a fit one
string is to another?

For example, if one string has one or more caps and the other doesn't

or one string has an "&" and the other has "and"

or one starts with "The" and the other doesn't

or one has a "2" and the other has "II"

etc.

Thanx in advance...
 
Those hardly constitute close fits, though, do they?

Sorry, I don't have anything like that.

Pete
 
Anyone have a canned subroutine that can determine how close a fit one
Hi. Not sure if this is what you want, but in Math, one option is the
"Levenshtein distance."

http://en.wikipedia.org/wiki/Levenshtein_distance

So, for example:

EditDistance["Kitten","sitting"]
3

EditDistance["Saturday","Sunday"]
3

EditDistance["Saturday","sunday",IgnoreCase->False]
4
 
Thanks for that, Dana - I'd never heard of it before. I wonder how one
would cope with the example the OP gave of one string containing 2 and
the other containing ii (or <space>ii<space>)?

Pete

Hi. Not sure if this is what you want, but in Math, one option is the
"Levenshtein distance."

http://en.wikipedia.org/wiki/Levenshtein_distance

So, for example:

EditDistance["Kitten","sitting"]
3

EditDistance["Saturday","Sunday"]
3

EditDistance["Saturday","sunday",IgnoreCase->False]
4

--
Dana DeLouis




Those hardly constitute close fits, though, do they?
Click to expand...
Sorry, I don't have anything like that.
Click to expand...

- Show quoted text -
Click to expand...
 
...one string containing 2 and the other containing ii.

Hi. Not sure on that one.
Anyway, such algorithms are used in different ways.
One example would be a function that links to a dictionary and uses that
"distance" algorithm as a default.

NearWords = Nearest[DictionaryLookup[__]];
Hence:

NearWords["Excel",10]

{"excel","Axel","Edsel","Ethel","excels","Exocet","expel","Abel","ace","aced"}

--
Dana DeLouis


Pete_UK said:
Thanks for that, Dana - I'd never heard of it before. I wonder how one
would cope with the example the OP gave of one string containing 2 and
the other containing ii (or <space>ii<space>)?

Pete

Anyone have a canned subroutine that can determine how close a fit one
string is to another?
Click to expand...

Hi. Not sure if this is what you want, but in Math, one option is the
"Levenshtein distance."

http://en.wikipedia.org/wiki/Levenshtein_distance

So, for example:

EditDistance["Kitten","sitting"]
3

EditDistance["Saturday","Sunday"]
3

EditDistance["Saturday","sunday",IgnoreCase->False]
4

--
Dana DeLouis




Those hardly constitute close fits, though, do they?
Click to expand...
Sorry, I don't have anything like that.

Anyone have a canned subroutine that can determine how close a fit one
string is to another?
Click to expand...
For example, if one string has one or more caps and the other doesn't
Click to expand...
or one string has an "&" and the other has "and"
Click to expand...
or one starts with "The" and the other doesn't
Click to expand...
or one has a "2" and the other has "II"

Thanx in advance...- Hide quoted text -
Click to expand...

- Show quoted text -
Click to expand...
Click to expand...
 
Thanx Dana.

Both of the algorithms mentioned are very close to what I need and a
very good starting point to resolving some of the other issues.


...one string containing 2 and the other containing ii.
Click to expand...

Hi. Not sure on that one.
Anyway, such algorithms are used in different ways.
One example would be a function that links to a dictionary and uses that
"distance" algorithm as a default.

NearWords = Nearest[DictionaryLookup[__]];
Hence:

NearWords["Excel",10]

{"excel","Axel","Edsel","Ethel","excels","Exocet","expel","Abel","ace","aced"}
Click to expand...
 
Don't know what happened to my earlier post, but this reponse was
right on the money.

Thanks, Dana
 
Back
Top