search

cdt15 / lingam

Python package for causal discovery based on LiNGAM.
https://sites.google.com/view/sshimizu06/lingam
MIT License
382 stars 60 forks source link

lingam with latent confounder #8

Open Wangt-CN opened 4 years ago

Wangt-CN commented 4 years ago

Hi, thanks for your work. Would this package support the lingam with latent confounder?

sshimizu2006 commented 4 years ago

Thanks. Not yet, but we are planning to add such lingam methods with latent confounder.

sshimizu2006 commented 3 years ago

Hi, we have added a method with latent confounders: https://github.com/cdt15/lingam/blob/master/examples/RCD.ipynb

kapibarasan-desu commented 2 years ago

Thanks for your work. BottomUpParceLiNGAM, RCD algorithm and LiNA algorithm are introduced as the lingam with latent confounder, which model is more appropriate to use?

sshimizu2006 commented 2 years ago

Hi, it depends on your assumptions. The three methods assume different assumptions. BottomUpParceLiNGAM assumes that only exogenous observed variables may be confounded. RCD allows that any observed variables are confounded. LiNA tries to discover causal relations of latent factors rather than observed variables.

kapibarasan-desu commented 2 years ago

Thank you for your reply! I understand it.