-
VARIO_PI: Given a variable and its causes in multiple contexts, discovers a partition of the contexts, showing where the causal mechanism is invariant and where it changes. Seecausal_mechanism_search.py -
VARIO: Given a variable observed in multiple contexts, discovers a context partition and the causal parents of that variable. Seecausal_variable_search.py. -
VARIO_G: For a set of variables observed in multiple contexts, discovers context partitions and causal parents of each variable. Seecausal_dag_search.py.
data, index_Y, true_dag, true_partitions = generate_data_from_DAG(n_nodes=5, n_contexts=5, n_samples_per_context=500,
partition_Y=[[0, 1, 2],[3,4]],
min_interventions=5,max_interventions=5, scale=False, verbose=True, seed=3)
# VARIO_PI
oracle_partition, oracle_score, _ = causal_mechanism_search(data, index_Y, true_dag.parents_of(index_Y), greedy=False, verbose=True)
# VARIO
estim_parents, estim_partition, estim_score, _ = causal_variable_search(data, index_Y, [n for n in range(n_nodes) if not (n==index_Y)], greedy_mechanism_search=False, verbose=True)
cd vario
pip install -e .
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent