search

py-why / dodiscover

[Experimental] Global causal discovery algorithms
https://www.pywhy.org/dodiscover/
MIT License
90 stars 18 forks source link

EPIC: Baseline causal discovery algorithms #76

Open robertness opened 1 year ago

robertness commented 1 year ago

This epic targets a variety of baseline causal discovery algorithms in dodiscover. The goal is to have a MVP as a discovery library.

U-Sahaj commented 1 year ago

Hi Robertness, I am new to DoWhy. But what a great Python package I must say! Do you mean to say until you complete your MVP, there is currently no automated way to automatically discover causal factors in dataset? I wish I can join you in this project but I have so much to catch up about DoWhy, arming with what little I know about Algorithmic Complexity I wish to add to this library. Thank you ! U

adam2392 commented 1 year ago

@U-Sahaj I'm not actually sure what Robert meant when he opened this issue.

But we have existing algorithms, but they are in a development stage. Feel free to use with caution.

robertness commented 1 year ago

Hi U-Sahaj! MVP in this context meant minimum viable causal discovery library. Nothing to do with computational complexity:)

U-Sahaj commented 1 year ago

Thank you both.

I am actually referring to the availability of Python code to automatically discover causal factors from the last price time series of a particular stock symbol. I am hugely inspired by https://www.nature.com/articles/s42256-018-0005-0 and of course the promise of DoWhy Python library as well.

Thank you once again.