[R] Latent models

Bi-Info (http://members.home.nl/bi-info) bi-info at home.nl
Wed Nov 22 19:02:17 CET 2006


Wrong button. The email went out too quick...

This might be the one, Brian.

The problem was, and is, that the data is composed of two interrelated
sets of variables, from two completely different ways of measurement,
which I have to connect through some "latent" structure. Most variables
are dichotomous / slightly scaled (1,2,3) and only one is numerically
scaled. It's not reasonable to assume a linear relation, because I
looked at the Rank linearity condition and it is not satisfied, probably
none of the relations is linear. The statistics book advised me to look
for latent models.

I'll take a look at these packages. Additional suggestions are welcome.



Prof Brian Ripley schreef:
> Those are only a few of the possible latent models.  Factor analysis 
> and SEM are also latent variable models, as indeed are the mixed 
> models of nlme, lme4, mmlcr, polr models ....  Then there are the 
> graphical models of (at least) packages ggm and latentnet, polytomous 
> LDA in package polLCA ....
> help.search("latent", agrep=FALSE) on a comprehensive installation 
> helped me find some of these.
> On Wed, 22 Nov 2006, ronggui wrote:
>> Is "ltm" what you want?
>>> packageDescription("ltm")
>> Package: ltm
>> Title: Latent Trait Models under IRT
>> Version: 0.6-1
>> Date: 2006-10-02
>> Author: Dimitris Rizopoulos <dimitris.rizopoulos op med.kuleuven.be>
>> Maintainer: Dimitris Rizopoulos <dimitris.rizopoulos op med.kuleuven.be>
>> Description: Analysis of multivariate dichotomous and polytomous data
>>       using latent trait models under the Item Response Theory
>>       approach. It includes the Rasch, the Two-Parameter Logistic,
>>       the Birnbaum's Three-Parameter, and the Graded Response Models.
>> Depends: R(>= 2.3.0), MASS, gtools, msm
>> LazyLoad: yes
>> LazyData: yes
>> License: GPL version 2 or later
>> URL: http://wiki.r-project.org/rwiki/doku.php?id=packages:cran:ltm
>> Packaged: Tue Oct 3 09:07:38 2006; dimitris
>> Built: R 2.4.0; ; 2006-10-03 17:16:04; windows
>> If you are looking function for "Latent Class Analysis (LCA)", then
>> "lca" in package e1071 is what you want.
>> On 11/22/06, Bi-Info (http://members.home.nl/bi-info) 
>> <bi-info op home.nl> wrote:
>>> Hai,
>>> Can anyone help me with some literature (R related) about latent 
>>> models?
>>> Thanx,
>>> Wilfred
>>> ______________________________________________
>>> R-help op stat.math.ethz.ch mailing list
>>> https://stat.ethz.ch/mailman/listinfo/r-help
>>> PLEASE do read the posting guide 
>>> http://www.R-project.org/posting-guide.html
>>> and provide commented, minimal, self-contained, reproducible code.

More information about the R-help mailing list