|Date: ||Thu, 7 Mar 1996 13:49:54 +0100|
|Sender: ||"SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>|
|From: ||Roberto Ferraro <FERRARO@IRFMN.MNEGRI.IT>|
|Subject: ||A test for sample distribution|
We need information related to mixture distributions and performing
test of goodness of fit with respect to a specific continuos variable,
measures of a diagnostic index for controls and patients.
We want to perform a test to verify if sample distribution is better
fitted by one normal than two normals or more.
Anybody knows if such test is possible to implement using a SAS proc ?
We would appreciate any informations.