Abstract
We propose a Goodness of Causal Fit (GCF) measure which depends on Pearl "do" interventions. This is different from a measure of Goodness of Fit (GF), which does not use interventions. Given a DAG set ${\cal G}$, to find a good $G\in {\cal G}$, we propose plotting $GCF(G)$ versus $GF(G)$ for all $G\in {\cal G}$, and finding a graph $G\in {\cal G}$ with a large amount of both types of goodness.
Abstract (translated)
URL
https://arxiv.org/abs/2105.02172
