[S] Bivariate Density Estimation

Wolfgang Rolke (w_rolke@rumac.upr.clu.edu)
Thu, 22 Oct 1998 11:22:26 -0700


I have a question concerning density estimation for
bivariate data. How do I judge the quality of the fit? My first
try was to use the Chi-square criterion but for technical reasons
that computation takes much to long. Are there any other useful
criteria? Is there a graph that I could use? In one dimension I
would just draw a histogram and the overlay it with the density
estimate. This does not seem feasible in two dimensions because
the graph becomes much to crowded.
The data I am working on is from High Energy physics. I
am trying to fit what is called a Dalitz plot. This is typically
a mixture of an almost uniformly distributed background and one or
two multivariate normals.
If anybody can point me to a good reference or if anybody already
has an Splus routine written I would be most grateful.


This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news