On Wed, Feb 28, 2001 at 12:26:49PM -0500, Bjorn H Samset wrote:
>
> With the ClusterVertex class, there has always been an issue of what cuts
> to use on the gauss-fit to select only "well-determined" vertices.
>
IMHO, you should convert your Chi-squared to a confidence level
and cut on the confidence level. As a bonus, you get to determine
your measurement resolution (otherwise the conf.level is not flat).
K.O.
> I presented my "best" values at the last collab.meeting and improved them
> somewhat before QM, but for diverse reasons they have never been made
> default in the class. Now that the class is beeing used by different
> people, I should long ago have checked this in so that the clustervertex
> may me used "blindly" - sorry about the slackness there. I have a version
> ready for checkin, but before I do so I want to ask if this messes up
> anything for anyone. What will happen is that instead of ClusterVertex
> returning a vertex whenever it is able to fit something to its data, it
> will only return a vertex when the fit is better than some
> (settable!) cuts. This will reduce the number of vertices but _greatly_
> increase the vertex precision.
>
> For reference, the planned default cuts are:
> BrVertex *cvtx = some_event_node->GetObject("Cluster VtxData");
> if (cvtx->GetVertexZSigma()<2.0 &&
> sqrt(cvtx->GetVarianceZ())<1.0) {OK!}
>
> If there are no comments, I'll check this in by the end of the week.
>
> Ping :-)
>
> ------------------------------------------------
> Bjorn H. Samset
> Master-student in Heavy Ion physics
> Mob: +47 92 05 19 98 Office: +47 22 85 77 62
> Adr: Kri 2A709 Sognsveien 218 0864 Oslo
>
>
--
Konstantin Olchanski
Physics Department, Brookhaven National Laboratory, Long Island, New York
olchansk@bnl.gov
This archive was generated by hypermail 2b29 : Wed Feb 28 2001 - 19:51:08 EST