nonlinear-causal 0.3

nonlinear-causal is a Python module for nonlinear causal inference built on top of Two-stage methods.

🧬 nonlinear-causal

nonlinear-causal is a Python module for nonlinear causal inference, including hypothesis testing and confidence interval for causal effect, built on top of two-stage methods.

• GitHub repo: https://github.com/nl-causal/nonlinear-causal
• PyPi: https://pypi.org/project/nonlinear-causal/
• Open Source: MIT license
• Paper: arXiv:2209.08889
• Documentation: https://nonlinear-causal.readthedocs.io

The proposed model is:

$$ \phi(x) = \mathbf{z}^\prime \mathbf{\theta} + w, \quad y = \beta \phi(x) + \mathbf{z}^\prime \mathbf{\alpha} + \epsilon $$

• $\beta$: marginal causal effect from X -> Y;
• $\phi(\cdot)$: nonlinear causal link;

What We Can Do:

• Estimate $\theta$ and $\beta$.
• Hypothesis testing (HT) and confidence interval (CI) for marginal causal effect $\beta$.
• Estimate nonlinear causal link $\phi(\cdot)$.

Installation

Install nonlinear-causal using pip

pip install nonlinear-causal

Install the latest version in Github:

pip install git+https://github.com/nl-causal/nonlinear-causal

Examples and notebooks

• User Guide

• Simulation for HT and CI with standard setup

• Simulation for HT and CI with invalid IVs

• Simulation for HT and CI with categorical IVs

• Real application

Reference

If you use this code please star 🌟 the repository and cite the following paper:

• Dai, B., Li, C., Xue, H., Pan, W., & Shen, X. (2022). Inference of nonlinear causal effects with GWAS summary data. arXiv preprint arXiv:2209.08889.

@article{dai2022inference,
  title={Inference of nonlinear causal effects with GWAS summary data},
  author={Dai, Ben and Li, Chunlin and Xue, Haoran and Pan, Wei and Shen, Xiaotong},
  journal={arXiv preprint arXiv:2209.08889},
  year={2022}
}