Re: ggplot2 for analyzing results of Mixture Models

[Dennis Murphy]

Hi:

I don't see anything that is exceptionally difficult - this should be straightforward to do in ggplot2. You need one aesthetic for color, another for shape; you'd also need an external character vector for the legend labels that you could pass into a scale_* function.

I'm a little confused about the legend, though: your graph has 12 separate lines (six colors, two shapes), but the legend only contains six merged color-shape combinations. By default, ggplot2 would produce two legends: one for the color aesthetic and another for the shape. 

Reproducible example?

Dennis

On Thu, Nov 22, 2012 at 11:48 AM, Andrey K. <koval....@gmail.com> wrote:

Hello everyone,

  i'm fitting latent growth curve models (lcgm) and general mixture models (gmm) and was wondering if anyone had some good templates to view the results of those analyses. I am working in Mplus, that produce charts like figur A, that shows the estimated trajectories for 6 latent classes, or figureB that overimposes observed individual trajectories over the class trajectories. Anyone used ggplots for these models?

--
You received this message because you are subscribed to the ggplot2 mailing list.
Please provide a reproducible example: https://github.com/hadley/devtools/wiki/Reproducibility
 
To post: email ggp...@googlegroups.com
To unsubscribe: email ggplot2+u...@googlegroups.com
More options: http://groups.google.com/group/ggplot2

[Brandon Hurr]

This sounds like something Josh Wiley has done. I could swear he is working on a package that connects R to Mplus as well. 

B