search

WillianFuks / tfcausalimpact

Python Causal Impact Implementation Based on Google's R Package. Built using TensorFlow Probability.
Apache License 2.0
600 stars 72 forks source link

Documentation #25

Open ghost opened 2 years ago

ghost commented 2 years ago

It would be really great if there was some good documentation on what are the possible arguments of the function(s) and how to use them.

For instance from the source code I see that the python version is behind the R package. In the R package, the causal impact function accepts another argument (dynamic regression) which here seems not available.

WillianFuks commented 2 years ago

Hi @msh855 ,

As for the input docs, probably the best source for now is the docstrings itself. There you'll find all available inputs and the explanation on how to use which and when.

As for the dynamic regression, it's also possible to use it in this package. Notice that for doing so you just need to build a customized model and use the dynamic component from the probability package.

I don't quite remember how this feature is implemented in the original R package, if it's something that can be specified in the input then we could open a new issue on this repo requesting this feature as well (probably I won't be able to implement that in the near term but if someone in the community wants to try it out it's totally welcome!).

Let me know if the docstrings are helpful to you.