<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=iso-8859-1">
<META content="MSHTML 5.50.4916.2300" name=GENERATOR></HEAD>
<BODY style="MARGIN-TOP: 2px; FONT: 8pt Tahoma; MARGIN-LEFT: 2px">
<DIV><FONT size=2>Dear R users,</FONT></DIV>
<DIV><FONT size=2></FONT> </DIV>
<DIV><FONT size=2>I am trying to find a package that can estimate</FONT></DIV>
<DIV><FONT size=2>an autoregressive </FONT><FONT size=2>model for discrete
data. I am </FONT></DIV>
<DIV><FONT size=2>imagining </FONT><FONT size=2>a Poisson or Gamma process in
which the </FONT></DIV>
<DIV><FONT size=2>mean </FONT><FONT size=2>(say mu) </FONT><FONT size=2>follows
a process such as</FONT></DIV>
<DIV><FONT size=2></FONT> </DIV>
<DIV><FONT size=2>mu_t = a + b*x + c*mu_{t-1}</FONT></DIV>
<DIV><FONT size=2></FONT> </DIV>
<DIV><FONT size=2>Suppose I have data on the time-series Poisson </FONT></DIV>
<DIV><FONT size=2>outcomes and x and </FONT><FONT size=2>would like to obtain ML
estimates </FONT></DIV>
<DIV><FONT size=2>for b and c.</FONT></DIV>
<DIV><FONT size=2></FONT> </DIV>
<DIV><FONT size=2>Does anyone know of a package that can do this or
</FONT></DIV>
<DIV><FONT size=2>something similar in R? My first (naive) instinct
was</FONT></DIV>
<DIV><FONT size=2>to use glm and the lagged outcome as a regressor,</FONT></DIV>
<DIV><FONT size=2>but that isn't quite right.</FONT></DIV>
<DIV><FONT size=2></FONT> </DIV>
<DIV><FONT size=2>Thanks very much,</FONT></DIV>
<DIV><FONT size=2></FONT> </DIV>
<DIV><FONT size=2>Michael Roberts</FONT></DIV></BODY></HTML>