ipd
R package
Inference with predicted data
Causal inference
Inference with predicted data for generalized linear models and beyond.
ipd is an R package implementing a unifying framework for inference with predicted data, including methods such as PPI, RePPI, and prediction-powered inference variants.
Key features:
- Unified interface that mirrors
glm() - Support for multiple IPD estimators via a single high-level API
- Tools for diagnostics, simulation, and workflow integration
- GitHub: https://github.com/ipd-tools/ipd
- Documentation: coming soon
- DOI / paper: https://doi.org/10.1234/ipd-package