How can I name the dimensions of my variables?

Questions v5
1

I’m running a multiple regression using pymc3 with many input variables. My betas look like this:

beta = pm.Normal(name="beta", mu=0, sd=100, shape=len(num_feats))

The shape of Beta depends on how many numerical input features i’m using. Is there any way to name the various levels? For instance if I have 3 numerical input feautres (temperature, pressure, frequency), can I name them as such instead of referring them to index 0, 1, and 2?

If I can’t do it directly in the distribution object I suppose that’s fine, but I also haven’t been able to find a way to plot the posteriors by referring to each dimension by name. They are always referred to as the “Nth” index of the distribution Beta.

Thanks for any help!

2

The names of the features would be the coordinates. Look at example codes where people have used coords= instead of shape= .

1 Like
3

You saved me. Thank you.