Hi again, To clarify what I mean, I specifically want to know the best way to degrade my confidence due to the fact that I'm missing data. Cheers, Denver. -- Denver Dash http://www.sis.pitt.edu/~ddash ----- Original Message ----- From: "Denver Dash" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Sunday, April 01, 2001 5:16 PM Subject: [UAI] confidence estimates for EM > Hi. > > Can anybody point me to the best reference for calculating confidence > intervals for parameters induced by the EM algorithm? Specifically I'm > talking about Bayesian networks, but other model families will probably do. > > Thanks in advance, > Denver. > > -- > Denver Dash http://www.sis.pitt.edu/~ddash >
