Why is simulateLeeHastie saving latent variables?

[samblechman]

I am using the simulateLeeHastie function to generate a ground truth DAG and synthetic data from it (20 measured and 5 latent variables, in this case). The ground truth DAG includes the latent variables, but marks them with parentheses, e.g., (X1);X2.

However, when I convert the tetrad data to pandas, the pandas dataframe sometimes (but not always!) contains some (or all, or none) of the latent variables. Why is this? Is there a way that I can ensure that latent variables do not end up in the tetrad data set?

Thanks!

[jdramsey]

Oh good point. I'm headed up a to a coffee shop in a few minutes; if I have a few minutes I'll take a look.

[samblechman]

Thank you for your quick response! It would also be nice to have some way of recording which nodes are latent (to visualize with a different color, for example). Is there some way to do this programmatically?

[jdramsey]

From Python?? In the Tetrad interface (i.e. the Java interface) they're represented differently. How are you plotting them?

[jdramsey]

Hmm... are you usign GraphViz?

[samblechman]

Yes, using GraphViz from python.

[jdramsey]

You can write a graphviz output however you like actually; it's very flexible. You don't need to use the one we provide. You can make the latent and measured variables appear however you want them to appear, and there are various canned graph layouts that are available.

[jdramsey]

Actually give me a few days to get back to you on all this--I have another project I also need to work on this afternoon as well...

[samblechman]

Okay! Thanks so much.