[mb-devel] Collaborative Filtering: Artist - Artist Relationships (Summer of Code Proposal)

[Jim C. Nasby]

On Thu, Mar 29, 2007 at 03:16:45PM +0530, sharon myrtle wrote:
> >What about this method:
> >keep two copies of the user rating count: the one used for the last
> >vector computation, and the current count. The first would be updated
> >only when the vector is computed again. The second would be updated
> >each time a user changes his vote. Then trigger the computation only
> >when the differences between the 2 counts is large enough to justify a
> >new computation. This could be done immediately (when the user changes
> >his vote) or by periodically browsing the counts. Of course this means
> >that you are able to decide what is the minimum difference which
> >should trigger a computation.
> 
> Yes, this would be a great way to get around the problem of excessive
> computation as well as be up-to-date with any major change in the rating
> values.
> 
> [actually, I implied something along these lines (of storing 2 copies) when
> I mentioned about calculating correlation vector and then updating the
> vector with the mean of the current ratings after a time lapse :-)]

No matter how you decide when to re-calculate ratings, you do *NOT* want
to do that in real-time. ACFs tend to generate a very large amount of
data and system load, so you want a way to control that.
-- 
Jim C. Nasby, Database Architect                decibel at decibel.org 
Give your computer some brain candy! www.distributed.net Team #1828

Windows: "Where do you want to go today?"
Linux: "Where do you want to go tomorrow?"
FreeBSD: "Are you guys coming, or what?"