d-separation in causal inference

written by Eric J. Ma on 2018-08-06

causal inference bayesian data science

Yesterday evening, I had an empty block of time during which I finally did a worked example of finding whether two nodes are "d-separated" in a causal graph. It was pretty instructive to implement the algorithm. It also reminded me yet again: there's this weird thing about me where I need programming to learn math!

Anyways, if you're interested in seeing the implementation, it's available at GitHub.