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 |
|
gavinharriss
Starting Member
17 Posts |
 Posted - 2008-04-08 : 21:01:13
|
| Hi,When using FREETEXTTABLE the RANK returned seems to be partially based on keyword density. Has anyone come across a method of compensating for this keyword density so that good matches found in a lot of text and a small amount of text return the same RANK?For an example look at the site I'm working on when someone searches for "whale watching" - http://www.yougodo.com/Search.aspx?ks=whale+watching - you can see that we are showing poor quality results at the top as our first sort order is based on RANK.If we could remove the keyword density factor from RANK then this would allow second order sorting criteria to come in to play and move the more valuable results to the top.Any help, pointers, advice would be greatly appreciated,Gavin.www.gavinharriss.com |
|
|
snSQL
Master Smack Fu Yak Hacker
1837 Posts |
Posted - 2008-04-08 : 23:22:59
|
| How about just dividing the rank by the length of the text? It won't be too sophisticated or take into account the word count, but it would be better than the simple rank. |
 |
|
|
|
|
|
Compensate for keyword density in Full Text Search