[CS241] Re: CS241 Digest, Vol 6, Issue 6

[Avram Levi]

Yuri sorry I was away all day... But it's great that you confirmed my 
numbers...

Why do we need to calculate P(path|VP0, VP1)? was that something that 
Eugene mentioned and I missed?

on the other hand, I calculated the P(VP1|VP0, path)/P(VP1|VP0) ratios, 
i think that's a pretty good metric that gives a lot of info...

See you soon,
Avram

Skip wrote:

> Why not P(path | VP0, VP1) = #(path, VP0, VP1)  / #(VP0, VP1) ?
>
> On 10/25/07, *Yuri Malitsky* <yuri.malitsky at gmail.com 
> <mailto:yuri.malitsky at gmail.com>> wrote:
>
>     Ok, by the time my plea for debugging info was posted, I already
>     solved the problem and confirmed Avram's numbers :/
>
>     Does anyone have any good ideas for calculating P( path | VP0 ,
>     VP1 )? I was thinking of something along the lines of:
>
>        alpha* ( #( path, VP0, VP1 ) / #paths ) + beta*( #( path, VP0,
>     VP1 ) / #(path) )
>
>     and then normalizing all the scores, but there has to be a better
>     approach.
>
>
>
>
>     _______________________________________________
>     CS241 mailing list
>     CS241 at list.cs.brown.edu <mailto:CS241 at list.cs.brown.edu>
>     http://list.cs.brown.edu/mailman/listinfo/cs241
>
>
>------------------------------------------------------------------------
>
>_______________________________________________
>CS241 mailing list
>CS241 at list.cs.brown.edu
>http://list.cs.brown.edu/mailman/listinfo/cs241
>  
>