grf-labs/grf

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1106 stars · 282 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

GRF is an R package (with C++ core) for forest-based statistical estimation and inference, providing non-parametric methods for heterogeneous treatment effect estimation, instrumental variables, and various regression types. It supports honest estimation, confidence intervals, and missing covariates.

Use cases

• estimate heterogeneous treatment effects from observational data
• fit a causal forest to measure treatment effect heterogeneity
• run quantile regression with random forests
• estimate treatment effects with instrumental variables
• perform survival regression with right-censored outcomes
• get confidence intervals for treatment effect estimates
• learn optimal policies from estimated effects

When to choose

• you need statistically valid inference (confidence intervals) for forest-based estimates
• you want non-parametric heterogeneous treatment effect estimation in R
• you need honest splitting and support for missing covariates
• you work in econometrics or causal inference with multiple treatment arms or IV designs

When to avoid

• you need a Python-native library rather than an R package
• you only need standard prediction without statistical inference
• you need deep learning or gradient boosting approaches instead of forests

Facets

library · maturity stable

machine-learning data-science machine-learning data-science cpp cross-platform causal-inference causal-forest random-forest econometrics heterogeneous-treatment-effects r-package statistics survival-analysis r

Member repositories

For agents

markdown · JSON · MCP: product_card(name="grf-labs/grf")