Changing encoding norms and boosting...

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Changing encoding norms and boosting...

escher2k
This is related to an earlier posting (http://www.nabble.com/Document-boost-not-as-expected...-tf3476653.html).
I am trying to determine a ranking for users that is between 1 and 1.5. Because of the way the encoding
norm is stored, if index time boosting is done, everyone gets a score of 1, 1.25 or 1.5. Is there any way
to get around this so that all the values can be retrieved as is (e.g. 1.22, 1.35 etc).

Thanks in advance.
Reply | Threaded
Open this post in threaded view
|

Re: Changing encoding norms and boosting...

Mike Klaas
On 3/29/07, escher2k <[hidden email]> wrote:
>
> This is related to an earlier posting
> (http://www.nabble.com/Document-boost-not-as-expected...-tf3476653.html).
> I am trying to determine a ranking for users that is between 1 and 1.5.
> Because of the way the encoding
> norm is stored, if index time boosting is done, everyone gets a score of 1,
> 1.25 or 1.5. Is there any way
> to get around this so that all the values can be retrieved as is (e.g. 1.22,
> 1.35 etc).

One option is to override the norm encoding in Similarity, and
re-index.  You'd still be limited to 256 values.

You could also store doc boosts that span a wider dynamic variance (1
to 15 rather than 1 to 1.5, say), then compensate by applying a
query-time boost of 0.1.

-Mike