From: Kristian M. <mar...@tc...> - 2004-04-15 19:08:55
|
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type"> <title></title> </head> <body> Baris-<br> <br> At one time, I had Brad's technical papers where he outlined the estimation procedure used in ICL. They were papers given at psychometrics conferences. <br> <br> I can't find them now, unfortunately, and his home page is no longer there.<br> <br> However, from what I remember of the estimation procedure, it is very similar--almost identical or actually identical--to the procedures used in the modeling program LEM. The approach taken by ICL is also essentially identical to that described by Heinen (1996).<br> <br> LEM is a program developed by J. K. Vermunt, and is freely available at <a href="http://spitswww.uvt.nl/%7Evermunt/">his webpage</a>. It is a general log-linear modeling program that includes IRT models; it's more flexible than ICL, but also somewhat more difficult to use in my opinion. It is also has the limitation of not being open source.<br> <br> The Heinen citation is<br> <br> Heinen, T. (1996). <i>Latent class and discrete latent trait models</i>. Thousand Oaks, CA: Sage.<br> <br> The ICL, LEM, and Heinen approach is similar to marginal maximum likelihood approach (MML), but the latent distribution is estimated rather than fixed. It can also be thought of as a type of latent class or mixture model. EM estimation is used.<br> <br> Hope that helps, <br> <br> Kristian Markon<br> <br> <br> <br> Baris Sahin wrote: <blockquote cite="mid000e01c41724$f6f63a90$c802000a@bs" type="cite"> <meta http-equiv="Content-Type" content="text/html; "> <meta content="MSHTML 6.00.2800.1400" name="GENERATOR"> <style></style> <div> </div> <div><font face="Arial" size="2">Dear all,</font></div> <div> </div> <div><font face="Arial" size="2">I need information about mathematical calculations used in your</font></div> <div><font face="Arial" size="2">project on sourceforge.net.</font></div> <div><font face="Arial" size="2">Item parameter estimation, vs...</font></div> <div> </div> <div><font face="Arial" size="2">Dou you have any documents, or bookmarks about that.</font></div> <div> </div> <div><font face="Arial" size="2">Thank you,</font></div> <div> </div> <div><font face="Arial" size="2">baris</font></div> <div><font face="Arial" size="2"><a href="mailto:bs...@mi...">bs...@mi...</a></font></div> <div> </div> </blockquote> <br> </body> </html> |