Please start any new threads on our new
site at https://forums.sqlteam.com. We've got lots of great SQL Server
experts to answer whatever question you can come up with.
Author |
Topic |
Thrasymachus
Constraint Violating Yak Guru
483 Posts |
Posted - 2005-08-16 : 15:44:07
|
If you know anything about a good machine aided indexing system I am looking for something like that as well. The one thing we evaluated only matched up at 30% with our data entry people. I am not even sure how to interpert those results. Seems so subjective.====================================================Regards,Sean RoussyThank you, drive through |
|
elwoos
Master Smack Fu Yak Hacker
2052 Posts |
Posted - 2005-08-17 : 03:27:25
|
Please excuse my ignorance but what is machine aided indexing?steveAlright Brain, you don't like me, and I don't like you. But lets just do this, and I can get back to killing you with beer. |
|
|
Kristen
Test
22859 Posts |
Posted - 2005-08-17 : 03:39:47
|
Would some sort of document management system do?Kristen |
|
|
Thrasymachus
Constraint Violating Yak Guru
483 Posts |
Posted - 2005-08-17 : 09:17:31
|
I am not sure if document management as I understand it would do. What we do now, is that we have subject matter experts using controlled vocabularies combing through volumes of material applying certain phrases and words to certain pages and documents.So it is not really key word or full text indexing. It is kind of semantic in nature. We have examined one product that can handle this. Still looking at the results.====================================================Regards,Sean RoussyThank you, drive through |
|
|
Kristen
Test
22859 Posts |
Posted - 2005-08-17 : 09:43:09
|
Ah ... more sort of Google than Document Mangement!!I can't begin to imagine how you solve that problem autoMagically."This product does NOT work with ABC123"Search on ABC123 .... shouldn't find this one ... but ... Kristen |
|
|
Thrasymachus
Constraint Violating Yak Guru
483 Posts |
Posted - 2005-08-17 : 09:50:30
|
Yeah but we are dealing with less data. Academic databases and the search results are more dead on because our controls are very tight and our links are verified. The whole thing is very expensive but somehow we maintain a profit.Anything machine aided is going to have too many rules in my opinion but we are looking at it anyways. Thanks for the feedback.====================================================Regards,Sean RoussyThank you, drive through |
|
|
Kristen
Test
22859 Posts |
Posted - 2005-08-17 : 09:55:08
|
Long ago a client had a system for keywording that I quite liked the look of - dunno if its applicable.Basically for each new "description" it "split" the words out and checked them against its own dictionary.It threw away the noise words, converted known-words via an "alias" table (so "Cars" became "Car" and so on), and then offered the cleaned-up known-words for any rogues to be explicitly excluded for this instance.Separately it offered any "new" words for consideration of adding to the Dictionary (including checking if they were aliases of existing words).All seems a bit straight forward to me, so I expect you are already way ahead of something that simple.Kristen |
|
|
spirit1
Cybernetic Yak Master
11752 Posts |
Posted - 2005-08-17 : 10:08:32
|
hmm... seems that this is something a well trained neural network might do...Go with the flow & have fun! Else fight the flow |
|
|
|
|
|
|
|