Expand ↗
Page list (1404)

c-Component

In a semi-Markovian causal model (an Acyclic Directed Mixed Graph), a confounded component — a maximal set of vertices connected to one another by paths of bidirected (latent-confounder) edges. The joint distribution factorises into “c-factors” over the c-components, and this decomposition is the engine of the ID Algorithm: identification is solved component-by-component, and non-identifiability is localised to particular c-components (see Hedge (Causal Inference)).

In this vault

Backlinks