Eta
Docs
std.causal.learn
Structure-learning helpers for causal graphs.
(import std.causal.learn)
Conditional-independence tests
Each test returns (independent? . p-value).
| Symbol | Description |
|---|---|
(learn:ci-test:fisher-z data x y z alpha) | Fisher-Z partial-correlation test. |
(learn:ci-test:chi2 data x y z alpha) | Discrete chi-square test. |
Search
| Symbol | Description |
|---|---|
(learn:pc data alpha ci-test) | PC algorithm: skeleton, collider orientation, Meek propagation. Returns a CPDAG with (u -> v) and (u -- v) edges. |
(learn:fci data alpha ci-test) | Latent-aware hook; currently delegates to learn:pc. |
(learn:ges data [alpha [ci-test]]) | Score-search hook; currently delegates to learn:pc. |
(learn:notears data lambda1 max-iter) | Continuous-relaxation learner using correlation thresholding plus transitive reduction. |