Am 10.07.2021 um 11:24 schrieb Riccardo (Jack) Lucchetti:
On Fri, 9 Jul 2021, riles(a)triton.net wrote:
> At the Gretl 2021 conference there was a discussion about adding
> Machine Learning and Causal Inference to Gretl. I was reading an
> article and it talked about EconML
> (
https://github.com/Microsoft/EconML) which is a python package from
> Microsoft that incorporates the three. The article can be found here
>
https://towardsdatascience.com/causal-inference-example-elasticity-de4a3e...
>
>
> Just thought I would share.
Aaaaah, this is _very_ interesting. Thanks for sharing!
Yes, thanks. As we already guessed or mentioned during the conference,
the main keyword there AFAICS is double machine learning = DML.
I do think there is some hype involved, though, but this may be my bias
coming mainly from macroeconomics, where I'm skeptical about fitting
non-parametric and/or nonlinear stuff to the short time series we
typically have. But in other areas of economics of course it could be
more promising.
thanks
sven