On the Identifiability of Poisson Branching Structural Causal Model Using Probability Generating Function

The Python implementation of the paper On the Identifiability of Poisson Branching Structural Causal Model Using Probability Generating Function (NeurIPS 2024).

Usage

The running example of PBSCM_PGF is given below.

from util import data_generate
from PB_SCM_PGF import PBSCM_PGF

parameter_dict = {
    "n": 10,
    "in_degree_rate": 3.0,
    "sample_size": 30000,
    "alpha_range_str": "0.1, 0.3",
    "mu_range_str": "0.05, 0.15",
}

data, edge_mat, _, _ = data_generate(**parameter_dict)
model = PBSCM_PGF(data.T)
model.learn()
causal_graph = model.dag

Requirements

The requirements are given in requirements.txt. You can install them using the following command:

pip install -r requirements.txt

Citation

If you find this useful for your research, we would appreciate it if you could cite the following papers:

@article{xiang2024identifiability,
  title={On the Identifiability of Poisson Branching Structural Causal Model Using Probability Generating Function},
  author={Xiang, Yu and Qiao, Jie and Liang, Zefeng and Zeng, Zihuai and Cai, Ruichu and Hao, Zhifeng},
  journal={Advances in Neural Information Processing Systems},
  volume={37},
  pages={11664--11699},
  year={2024}
}