Reply
 
Thread Tools Display Modes
  #1   Report Post  
Dave Williams
 
Posts: n/a
Default Fuzzy logic matching in a DLL

Hi all,

I'm developing an application that maintains a store of Word documents, and
one of the features is for basic searching into the docs. However, I'd like
to enhance it to allow a more 'clever' searching - heuristic or 'fuzzy
logic' or whatever it's called - to be able to judge what the document is
'about' or how similar it is to another document or phrase or something.

I'm ideally looking for a DLL or ActiveX object or something that I can call
in my app, giving it the rule of what to look for and the document, and it
will give me back a fitness rating for the doc, which I can then present to
the users in order of fitness. Preferably something with no UI of its own.

Can anyone recommend such a DLL or ActiveX product? Or am I approaching this
the wrong way?

Thanks,
Dave


  #2   Report Post  
Word Heretic
 
Posts: n/a
Default

G'day "Dave Williams" ,

There is no such dll. However, the basic logic is compare
ever-increasing fragment sizes of each string. Recent research
indicates that fragments beyond 4 letters long have reduced
effectiveness for comparison time taken.

Steve Hudson - Word Heretic

steve from wordheretic.com (Email replies require payment)
Without prejudice


Dave Williams reckoned:

Hi all,

I'm developing an application that maintains a store of Word documents, and
one of the features is for basic searching into the docs. However, I'd like
to enhance it to allow a more 'clever' searching - heuristic or 'fuzzy
logic' or whatever it's called - to be able to judge what the document is
'about' or how similar it is to another document or phrase or something.

I'm ideally looking for a DLL or ActiveX object or something that I can call
in my app, giving it the rule of what to look for and the document, and it
will give me back a fitness rating for the doc, which I can then present to
the users in order of fitness. Preferably something with no UI of its own.

Can anyone recommend such a DLL or ActiveX product? Or am I approaching this
the wrong way?

Thanks,
Dave


  #3   Report Post  
Dave Williams
 
Posts: n/a
Default

Thanks for that. It's disappointing if there's no purchasable products that
offer 'more intelligent' searching through Word documents, but thanks for
the suggested technique anyway.

"Word Heretic" wrote in message
...
G'day "Dave Williams" ,

There is no such dll. However, the basic logic is compare
ever-increasing fragment sizes of each string. Recent research
indicates that fragments beyond 4 letters long have reduced
effectiveness for comparison time taken.

Steve Hudson - Word Heretic

steve from wordheretic.com (Email replies require payment)
Without prejudice


Dave Williams reckoned:

Hi all,

I'm developing an application that maintains a store of Word documents,

and
one of the features is for basic searching into the docs. However, I'd

like
to enhance it to allow a more 'clever' searching - heuristic or 'fuzzy
logic' or whatever it's called - to be able to judge what the document is
'about' or how similar it is to another document or phrase or something.

I'm ideally looking for a DLL or ActiveX object or something that I can

call
in my app, giving it the rule of what to look for and the document, and

it
will give me back a fitness rating for the doc, which I can then present

to
the users in order of fitness. Preferably something with no UI of its

own.

Can anyone recommend such a DLL or ActiveX product? Or am I approaching

this
the wrong way?

Thanks,
Dave




  #4   Report Post  
Word Heretic
 
Posts: n/a
Default

G'day "Dave Williams" ,

There are bound to be third party apps that will do it on a file
search baiss for you - just nothing I know about within the std
implementation.

Steve Hudson - Word Heretic

steve from wordheretic.com (Email replies require payment)
Without prejudice


Dave Williams reckoned:

Thanks for that. It's disappointing if there's no purchasable products that
offer 'more intelligent' searching through Word documents, but thanks for
the suggested technique anyway.

"Word Heretic" wrote in message
.. .
G'day "Dave Williams" ,

There is no such dll. However, the basic logic is compare
ever-increasing fragment sizes of each string. Recent research
indicates that fragments beyond 4 letters long have reduced
effectiveness for comparison time taken.

Steve Hudson - Word Heretic

steve from wordheretic.com (Email replies require payment)
Without prejudice


Dave Williams reckoned:

Hi all,

I'm developing an application that maintains a store of Word documents,

and
one of the features is for basic searching into the docs. However, I'd

like
to enhance it to allow a more 'clever' searching - heuristic or 'fuzzy
logic' or whatever it's called - to be able to judge what the document is
'about' or how similar it is to another document or phrase or something.

I'm ideally looking for a DLL or ActiveX object or something that I can

call
in my app, giving it the rule of what to look for and the document, and

it
will give me back a fitness rating for the doc, which I can then present

to
the users in order of fitness. Preferably something with no UI of its

own.

Can anyone recommend such a DLL or ActiveX product? Or am I approaching

this
the wrong way?

Thanks,
Dave




Reply
Thread Tools
Display Modes

Posting Rules

Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
seeking template containing numbered heads matching para style sue-d Microsoft Word Help 7 November 29th 09 11:04 PM
Why are my toolbar icons "fuzzy"? Golfer21 New Users 2 January 14th 05 04:51 PM
Fuzzy text Chris Microsoft Word Help 3 January 13th 05 05:02 PM
How do I delete all non matching lines in a word file? Carl Microsoft Word Help 1 January 11th 05 07:16 PM
matching tables Cath Tables 3 November 11th 04 02:37 AM


All times are GMT +1. The time now is 03:54 PM.

Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Copyright ©2004-2024 Microsoft Office Word Forum - WordBanter.
The comments are property of their posters.
 

About Us

"It's about Microsoft Word"