lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mark Miller (JIRA)" <>
Subject [jira] Commented: (LUCENE-1896) Modify confusing javadoc for queryNorm
Date Mon, 07 Sep 2009 21:48:57 GMT


Mark Miller commented on LUCENE-1896:

Suggestions? I am no IR guy.

It would appear to me that this component is simply the part where you convert the vectors
to unit vectors. Or are you are just dividing by the product of the euclidean lengths - it
appears to be the same in either case to me.

So it looks like we kind of factor that out as a constant (I barley understand myself too
- don't worry) -

In which case, why do we need it? Is it just there as a reminent of the math? Yes, it will
give us the sim measure as the cosine - by why do we care if its a scaled cosine (eg not the
cosine, but the same *relative* scores) for less work? What do we need 1/(Sum(W^2))^1/2 for?
 Seems like wasted calculations. Who is the academic that kept this in! ... :)

Actually, I don't have a clue if it makes sense to keep it or not. Or even whether or not
I am talking out my ...

Wherefore art thou Doug Cutting.

> Modify confusing javadoc for queryNorm
> --------------------------------------
>                 Key: LUCENE-1896
>                 URL:
>             Project: Lucene - Java
>          Issue Type: Improvement
>          Components: Javadocs
>            Reporter: Jiri Kuhn
>            Priority: Minor
>             Fix For: 2.9
> See
> The javadoc confuses me as well.

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message