cdt15 / lingam

Python package for causal discovery based on LiNGAM.
https://sites.google.com/view/sshimizu06/lingam
MIT License
380 stars 59 forks source link

Prior adjancency matrices #4

Closed fccoelho closed 4 years ago

fccoelho commented 4 years ago

Why are prior adjacency matrices, created with the make_prior_knowledge function, have -1 for the edges instead of 1? what is the rationale for this?

ikeuchi-screen commented 4 years ago

In make_prior_knowledge, the edge that no prior knowledge is given is -1.

Prior knowledge is defined as the following [ref]:

The above explanation is missing in the API reference and tutorials, so it will be added in the next version.

fccoelho commented 4 years ago

I was guessing that was what it meant. Thanks for the clarification. BTW, thanks for putting this package together!