Jason K. Moore
Jason K. Moore
I would guess that omega_a = theta_a.diff(t).
Are you saying he fits multiple models to one set of measurements, with each model having different noise? I think I was interpreting it to multiple measurements (his episodes being...
I think he likely did.
In equation 5:  I would think that the sum over M in the objective would somehow indicate that this is a sum over the same measurements, i.e. the argument...
Ah, I see what you are saying. The $\theta_m(t)$ in the objective does not have an $s$ subscript. So it is the same for all M episodes.
Most all of the attributes are generated on instantiation of the class, so they will need be manually added to a list of attributes in the class's docstring.
Yes, but you may want to read the numpydoc format. You'd probably add info about attributes to the primary docstring of the class. See: https://numpydoc.readthedocs.io/en/latest/format.html#documenting-classes
Note the difference in an attribute and a method. Only document the attributes that do not start with `_`, i.e. the only the public attributes.
We should consistently document all public attributes. A little redundancy doesn't hurt.
> I had forgotten, Timo told me again. Should this be explained? You could explain that in the Problem documentation.