Statistical Causal Inference

Production-ready causal attribution and inference algorithms for high-performance applications.

Overview

This package provides the core algorithms and statistical methods for causal inference, used by the CausalMMA SDK and other applications.

Features

• Statistical Causal Inference: Advanced algorithms for causal effect estimation
• Causal Discovery: PC Algorithm, FCI, and other structure learning methods
• Optimized Performance: Vectorized operations with NumPy and Numba acceleration
• Async Processing: Efficient asynchronous computation with Dask
• LLM Integration: OpenAI integration for causal reasoning
• Production Ready: Comprehensive error handling and validation

Installation

pip install statistical-causal-inference

From source

git clone https://github.com/rdmurugan/statistical-causal-inference.git
cd statistical-causal-inference
pip install -e .

Usage

from causalinference.core.statistical_inference import StatisticalCausalInference, CausalMethod
from causalinference.core.statistical_methods import PCAlgorithm
import pandas as pd

# Statistical causal inference
sci = StatisticalCausalInference()
result = sci.estimate_causal_effect(
    data=df,
    treatment='treatment_column',
    outcome='outcome_column',
    method=CausalMethod.DOUBLY_ROBUST
)

# Causal discovery
pc = PCAlgorithm()
dag = pc.learn_structure(data=df)

Core Modules

• statistical_inference.py - Causal effect estimation methods
• statistical_methods.py - PC Algorithm and other statistical methods
• causal_discovery.py - Structure learning algorithms
• optimized_algorithms.py - Performance-optimized implementations
• async_processing.py - Asynchronous computation utilities
• llm_client.py - LLM integration for causal reasoning

Requirements

• Python >= 3.9
• NumPy >= 1.21.0
• Pandas >= 1.3.0
• Scikit-learn >= 1.0.0
• NetworkX >= 2.6.0
• SciPy >= 1.7.0

Development

# Install development dependencies
pip install -e ".[dev]"

# Run tests
pytest tests/

# Format code
black causalinference/

License

Proprietary - For use in CausalMMA projects

Contact

For questions or support, contact: durai@infinidatum.net

Version

Current version: 4.4.0